Studies On The Samaveda

STUDIES ON THE SMAVEDA
PART I
BY
BAREND FADDEGON
VERHANDELINGEN DER KONINKLIJKE NEDERLANDSE

AKADEMIE VAN WETENSCHAPPEN, AFD. LETTERKUNDE
NIEUWE REEKS DEEL LVII, N. 1
SECOND PRINTING: 1963

NO-RTH-HOLLAND PUBLISHING COMPANY
(N.V. Noord-Hollandsche Uitgevers Mij.)
AMSTERDAM
TABLE OF CONTENTS
INTRODUCTION
page
Read
at a meeting of tke Koninklijke Nederlandse Akademie van

Wetenschappen . . . . . . . . . . . . . . . .
The edition of the Sm.veda-Sarphit, 9; publications by R. Simon and
J. M. van der Hoogt, 9.

The indications of pitch, 9; uncertainty of their interpretation, 10; the pitch
of the musical tones and the accents of spoken language, 11; proposalof a
fictitious performance of the smans for the sake of further research, 11;
pentachords and hexachords arrived at by quinta.l tuning, 11; difference
between a melodious major third and a major third in equal tempering made
audible on the violin, 12; the term udha, 13; antiquity ofthe drone-tone, 13;
drone-tone to he made audible on an instrument, 14. The musical metre resting
on the poetical metre; duration of the tones, 14. The spiritual contents ofthe
chants, 15.
The exposition of details. lts aim the examining of the relation of the
spoken and chanted Vedic language, 15; (fictitious) way of making the bh~
accentuation audible, 15; !;tV. 1,1 recited according to the bh~-accentuation,
16; according to Smavedic accentuation, 17; according to !;tgvedic accentuation, 18; the Smavedic melodization of the same strophe, 18. Plan of
further exposition, 19.
1. Stobha-clauses, which placed in a sman keep the melodious movement
of the spoken language except in their omamental closing, 19; stobha SV. 1,
1321, p. 20, stobha SV. 1,511 14, p. 20. 11. A stobha-line and a sman following
in their melody the bh~-accentuation, however with exception that the
curved aftertone is mostly replaced by tone 1; stobha SV. 1,9()l, p. 21,
sman SV. 2,212, p. 21. 111. Smans for which sacral monotony is characteristic, 22; the sman to Savitar, 22; a sman of the h3'agna (2, 1, 3), p. 22.
(For a fourth form of melodization in which the pdas of the strophe are broken
up into three- or four-syllabic parvans, see Introductory Note to Chapter 11,
p. 58). Discussion on the SubrahmaQy, 23; its ritual use and crude popular
contents, 23.
How was music bom? 24; the smans as expression of religious emotion, 24.
CHAPTERS I AND II
How to 8tudy the Smaveda-gnas
25
BIBLIOGRAPEUCAL TABLE
26
CHAPTER I
The Vedic and clas8ical Hindu music witk 8pecial retere,nce to the
tonal sy8tem . . . . . . . .
. . . . . . . . . .
27
PRELIMINARY NOTE
The Smavedic gamut according to Burnell and Se!i!&giri Sstri.
27

SECTION I
The general line in the development of music
page
28
The Greek and Mediaeval European gamuts, 29; the discovery of physical
vibrational fra.ctions as the basis of psychical musical interva.ls, 30; the
expression of physical vibrational fra.ctions as intervals by logarithmical
calculation ; the notion of cents, 30; the flageolet-tones of the violin, 31;
the partial tones of the hom, 32; the Hindu sruti-system, 33; the sa.-gr,ma
in relation with the sruti-system, 34_
SECTION Il
HiBtorical evidence . . . .
35
THREE PRELIMINARY NOTES
35
1. The indication of the changes in the text when melodized, 35;

Il. Method a.ccepted for the transcribing of the a.ccentuation, 35; lIl. Method
adopted for transcribing the melodization, 35.
DOCUMENTS OF HISTORICAL EVIDENCE
36
SaIphitM, 36; the types, i.e. parvan-melodies mentioned by Simon, 36 br.hmaJ}.as, the terms upag.tar and sth,na, 38; - stras and commentaries,
39; passages in the Pu!Jpastra of special interest with reference to the fixation
of intervals, PS. 8,90; 9,30, p. 40; table on the treatment of the final syllable
in an ,rcika or staubhic nidhana in its spoken form, as yoni and as ha, 41;
PS. 8, 190, p. 44; the notion of type a.ccording to the Pu!Jpastra and Fox
Strangways, 45. - S~M, p. 47.
SECTION III
Modem tradition. . . . .
47
Fox Strangways' enthusiasm and ina.ccura.cy, 49; information given by

Burnell and Se~giri S~tri, 47; the author's hypothesis about the S,ma.vedic
scale; 2-flat, 50; modern way of reciting the ~gveda, 51; Burnell's information
about how the notes of the sa.-gr,ma are classified; the sa.-gr.ma possibly
developed from an original hexa.chord, 53; modem performance of a s.man
by a s,maga; comparison with the S.maveda-tradition, 53.
CHAPTER 11
Three analytical studies on the Siimaveda-Sarp,hitii
57
PRELIMINARY NOTE
Types- of melodization, 57; rules valid when the melody is ba.sed on the
a.ccentuation, 57; cases possible when the pooa is broken up into three- or
four-syllabic parvans, 58.
SECTION I
Olassijication of the chants of the Griimegeyagiina .
58
PRELIMINARYNOTES
J. General remarks; 11. Scheme of subdivision in the following

subsections
. . . . . . . . . . . . . . . . ..
58
STUDIES ON THE SAMAVEDA
Subsections I-IV
The tones 1-4 are final tones
Conclusion to Subsection I
5
page
58
60
Subsection V
Tone 5 is the final tone .
61
A. The final parvan consists of text..:syllables or a sentence-stobha; a. the

tone is a pralq1;i-tone, 61; b. the tone is a vikrti-tone, 62; B. the parvan
consists of or ends in a Bound-stobha, a. the sound is monosyllabic, 65;
b. the sound is bisyllabic, 66.
Subsection VI
Tone 6 is the final tone
67
GENERAL CONCLUSION TO THE SIX SUBSECTIONS
Function of the different tones of the gamut as final tone
67
SECTION 11
Analysis of the smans ending in 5656
Subgroup I /123455656/ . .
Subgroup 11 /23455656/' . . . .
68
69
Cla.ssification in tabellic form, p. 70.
Subgroup 111 /323455856/' . . .
74
Cla.ssification in tabellic form, p. 84.
SECTION 111
Oldenberg's rule of acoustic metalepsis. . .
77
Oldenberg's examples from the AraQyegeyagna (Ohyagna), 78; from the

Grmegeyagna, 79; examples added to those of Oldenberg from the
Ar8I:).yegeyagna, 80; from the Grmegeyagna, 81.
-,
POSTSCRIPT
Date of manuscript being finished
83
INTRODUCTION
READ IN THE NETHERLANDISH ORIGINAL
AT THE MEETING OF THE KONINKLIJKE
NEDERLANDSE AKADEMIE VAN WETEN.
SCHAPPEN ON THE Ilth OF OCTOBER 194,8
Chant and recital in the Veda - a theme belonging both to the history
of language and the history of music! Of this Veda the J;tg-Veda-SaIphita.
will only in passing ask our attention, of much greater interest will be the
Sa.ma-Veda-SaIphita., which, consisting for the greater part of J;tg-Vedic
verses, yet as a collection of sacred chants is provided with musical
notations. Moreover a passage in Pa.IPni's Aftltdhyyi, this wonderfully
concise grammar of accurate secular Sanskrit with incidental notes on
sacerdotal Vedic will be of prime importance.
The rks of the J;tg-Veda were recited by hotrakas and the sa.mans of
the Sma-Veda sung by chandogas, but before these chanters gave their
performance they had the opportunity of calling to memory completely
and accurately words and meanings of their chants. And this was very
necessary the hymns being sung so as to render everything quite unrecognizable. In these prayers of memorization, japas, the chanters undoubtedly will have used the accentuation in which during their preparatory education they had learnt the texts. This Smavedic accentuation
differed much from the language of daily life, and the modern reader gets
the impression of having to do here not only with a sacral convention, but
in one interesting rule with a dialectical peculiarity not mentioned in the
grammatical manuals that have come down to us.
Besides hotrakas and chandogas I shall have to mention at the end
of this introduction the adhvaryu, who in the recital of his formulae and
verses was bound to the same regulations as the hot'rakas.
Of the ~g-veda-SaIphita. we possess Max Mller's famous edition ; the
Sma-Veda-SaI!1hita. has been published by Satyavrata Smasrami in the
Bibliotheca Indica; this at all events is the most accessible edition ; it is
a work of great value, but by its large number of inaccuracies a source of
continual complaint. Theoretical works on the Vedic chants were translated
by Richard Simon. J. M. van der Hoogt has in his academical thesis
(Amsterdam 1929) shown how much research is still to be done before
we can quite realize the importance of Simon's learned and valuable
work.
Only for some of the main points I shall invite your attention. The
edition of the Bibliotheca Indica rests on one tradition, that of the Kauthumas, a sacerdotal school spread in recent times over the Gujara.t. They
indicated the pitch of the syllables by means of numerals, written for the
main tone on the top of the syllable, for the following tones either after
the vowel or in the case of split diphthongs such as -i and -u after
the fust component. The numerals go from 1 to 6 as an indication of a
descending gamut. IJoreover we meet with 11 and 7; 11 only twice for a
tone higher than 1, and 7 over syllables that begin on 2 and soon rise to 1
10
in fixed proportion of time; as the use of this gracenote is bound to the

condition of the preceding syllable being sung on tone 1, we have to do
here with a means of marking a new beginning. Next to these numerals
we meet the avagraha as an indication of pitch; when used in this meaning
it indicates the succession 12 likewise in prescribed proportion : aso called
vinata; in the Bibliotheca Indica it is followed by the superfluous (1)
numeral 2 or a zero, which perhaps indicates a longer sustaining of the
second component.
Which pitches were indicated by the numerals 1-6, is still uncertain.
The first information about this question was given by A. C. Burnell in
the year 1876. In his introduction to the text of a Smavedic BrhmaI,la,
then published, he informs us that he had the opportunity of listening to
the chants of the Kauthuma priests and of ascertaining by means of a
'standard pitch pipe' (p. xlii no. 1) tone 1 to be t, tone 2 to be e etc. in
diatonical order without accidentals. This then means that the numerals
1-6 are notations for two minor thirds divided from the top into a semitone and a whole tone and following each other at the distance of a fourth.
In the year 1877 the same author in his introduction to another Smavedic BrhmaI,la quoted four lines from a Hindu treatise on the theory
of music (Nrada-sikf:1, prap. 1 kh. 5, sI. 1 sqq.), which, as he believed,
confirm his experience with the Kauthama priests and the standard
pitch pipe.
The two passages in Burnell's introductions were not followed by aily
further information until the years 1901-1906, when the Descriptive
Catalogue of the Sanskrit Manuscripts in the Government OrientalLibrary
at Madras, a work of the librarian Prof. Sef:1agiri Sstri, was published
after the author's death by a colleague ofhis. In this catalogue (p. 76-78)
a description of the Smavedic gamut is given; the author correctly
translates the Nrada-sikf:1-passage, not well understood by Burnell, and
interprets it according to usage as meaning the tones t es dca bes, here
then the second and sixth tones are flattened, and the tones 4-6 do not
I;
11
form a regular descending line. In the chapter The Vedic and classical
Hindu Music I shall adduce arguments for my supposition that the oldest
form of the Smavedic scale was a pentatonic gamut arrived at by quintal
tuning, for instance a g ede or e d c. a g without a note 6, and that only
at the time ofthe Hindu sruti-system it was transformed. At all events
11
we have no trustworthy tradition about the original significance of the

Smavedic notations and we must be satisfied with hypotheses.
It may seem strange that for centuries the large bulk of J;tg-Vedic
texts were handed down by oral teaching with incredible accuracy, but
that there is absolute uncertainty about the intervals of the Smavedic
chant. Here then we have to consider two circumstances: 1 ly, the chants
of the Vedic times are closely connected with the accents of the spoken
language, and this accentuation has in course of time totally changed
its character; no longer is the unity of the word effected by one of the
syllables bearing an udtta, but one of its syllables being pronounced
with greater use of expiration and increased loudness.
21y, music and song in the classical period began to rest on a different
tonal basis. No longer were the tones of the tonal series fixed by tuning
in fifths with perhaps here and there an incidental correction, but the
whole range of tones was attained by a systematical tuning in fifths and
thirds. In consequence of this system a melodiously schooled Hindu
singer will consider the thirds and sixths of European music unsatisfactory and he will prefer the wealth of his own melodie intervals as an
expression of his emotions.
Of great consequence must have been the effe cts of the described
evolution for the tradition of the Vedic chant. On the one hand the Vedic
chanters cannot yet have known the differences 'fine as ahair', kaiBika,
characteristic for the classical tone-system, and on the other hand the
classica.! theoreticians and singers must have been puzzled at the antique
Vedic chants. Without doubt these chants were modernized in numerous
ways and even the gamut has been forgotten.
Now in view of the great number of questions which arise about the
ancient Hindu chants, for which questions the width of the intervals is
rather irrelevant, we have to choose a gamut more satisfactory than those
recommended by Burnell and Se~giri Sstri, for only in this way shall
we be able to commit the chants to memory for the sake of comparison
and analysis. The gamut, then, which I have chosen, is of a very old type;
it is one of the five pentatonie gamuts. In order to elucidate this point
let us take a series of five notes tuned in successive fifths, e.g. the strings
of the viola and violin:
9= J
c
r
g
6J
d'
If we place these tones in close succession :
ede g a c' d' e' g' a' etc.
a'
e"
12
we get a tonal series with intervals of either a whole tone or a minor

third. Out of this series we can form five pentachords in accordance
with the choice of the beginning tone: either c or d etc. These gamuts are
not equally satisfactory, because the quintal tuning gives rise here and
there to false thirds or sixths between the tonic and one of the following
tones (for instance c-e, c-a); another difference between the five gamuts
arises from the fact that two of these gamuts contain only one minor
third: edelga and galede, while the others have two leaps and consequently
a greater compass : delgalc etc. It seems to me that the gamut galc' d' e'
deserves preference, since it easily allows the addition of as sixth tone.
At all events I have chosen this gamut galc'd'e', in the transposed form
abd'e'(fis)', also for a practical reason: the relative position of the range
of my voice to the violin by which I support my intonation.
If we examine the two gamuts edega and gaede more critically, then
we meet in the former the major third c-e p.,nd the major sixth c-a, and
in the latter the major sixth g-e which for a melodiously trained ear are
not satisfactory, they are too wide in consequence of the quintal tuning.
In order to explain this briefly, let me say this much: since about two or
three centuries we have followed in Europe what is called the equal
tempering. In this method of tuning the octave is divided into twelve
semitones of exactly equal width, and all the other intervals consist of a
whole number of such semitones. The scientists who occupy themselves
with the study of acoustics have gone further in this direction : they
divide theoretically these equal semitones into a hundred equal parts
which they caU cents. With the aid of an easy calculation it is possible to
express any given interval in a sum of such cents. So it appears that the
quintal major third contains 408, the equally tempered major third 400
and the melodious major third 386 cents. For the major sixth the numbers
respectively are 906, 900 and 884. For the modern European, who is
accustomed to the equally tempered intervals, it is a mere question of
IV
corda
IV corda
differences fine as ahair, kaiika. The differences between the equally

tempered major third of 400 and the melodious major third of 386 cents,
and the equally tempered major sixth and the melodious major sixth
can be made audible by the violinist; in the former cases he uses the
firm finger pressure and in the latter cases he plays the tones as
flageolet-tones on the g-string.
The difference between the old Vedic and the classical Hindu tonal
systems, then, amounts to this: the musicians of the Vedic times fixed
the tones quintally, perhaps here and there correcting a third and sixth
13
by auditive intuitio.n. But in the classical perio.d a fixed system o.f tuning
was attained, with use o.f fifths, o.f melo.dio.us sixths and thirds; and this
system penetrated even in the fo.lkso.ngs. Similarly as in Greece Pythago.ras and his scho.o.l carefully studied the arithmetical pro.po.rtio.n between
the vibratio.ns o.f the musical intervals, so. in India a minute auditive
study o.f the intervals must have been made.
No.w we kno.w that Pythago.ras preached metempsycho.sis at abo.ut the
same time when in the Upani~ thc saIpsra was pro.pagated, that the
same Greek philo.so.pher and scho.lar laid do.wn the fo.undatio.n o.f geo.metry
when the theo.rem called after him was in India practically applied acco.rding to. the Sulvastras in the co.nstructio.n o.f altars. By analo.gy we may
fo.r the present surmise that the minute study o.f musical intervals in India
to.o.k place at abo.ut the end o.f the Vedic perio.d. The co.nsequence o.f this
study was that the o.riginally very simple pentato.nic scale was gradually
replaced by the intricate ruti-system with its quintal, tertial and septimal
tuning and its great variety o.f grmas, murcha1UUJ, rgaIJ etc.
It is therefo.re quiw o.ut o.f the questio.n that mo.dern Hindus educated
in their ruti-system co.uld sing a sman co.rrectly in the histo.rical sense
o.f the wo.rd; indeed their pro.ductio.n o.f smans sho.ws an indefinite numher
o.f perso.nal differences and is in flagrant co.ntrast with the no.tatio.ns o.f
the Smavedic manuscripts, which does no.t preclude that their rendering
may have in so.me cases a great charm, even fo.r an Occidental audience.
The gamut, then, which o.n histo.rical gro.unds and fo.r the sake o.f further investigatio.n I have accepted, runs in downward directio.n as edcag =
12345. But in a small set o.f smans, there was intro.duced a sixth
to.ne, which, ho.wever, did no.t get the character o.f a tonic. So perhaps
the pentacho.rd edcag was turned into. the hexacho.rd edcagl with to.ne 5
functio.ning as dro.ne tone and regular final tone. Further as an ancient>
so.urce info.rms us, the to.nes 1-3 were respectively uduluuJ o.r harmo.nic
derivatives o.f the tones 4-6; and in the practice o.f the chando.gas the
to.nes 4 and 5 were o.ften used as substitutes fo.r land 2.
As to. the use o.f tone 5 as dro.ne tone I sho.uld like to make the fo.llo.wing
remark. Acco.rding to. BrhmaJ}.a-texts there were in great sacrifices throo
main chanters who. partly sung an o.wn divisio.n o.f the chant, but partly
sung to.gether (tho.ugh uniso.no.) refrainlike parts, nidha1UUJ. The so.lo. so.ngs
o.f these chanters were acco.mpanied by three upagtars, subchanters,
who. sung the syllable ho o.n to.ne 5, and by the yajiiapati, who. sung the
ho.ly syllable om o.n the same pitch. One will no.tice that there is in Burnell's
gamut the distance o.f a diminished. fifth between tone I o.r land tone
5 o.r b. It is inco.nceivable that aVedie singer co.uld into.ne this to.ne I
co.rrectly and sustain it, as wo.uld o.ften he demanded, during a who.le
pda against the dro.ne to.ne 5. In the pentatonic gamuts such a difficulty
co.uld never o.ccur.
So far I have mentio.ned the facts which plead fo.r my hypothesis.
I must, ho.wever, co.nfess that there is a detail which sooms to. indicate that
at a rather early date there existed between the tones 1 and 3 a minor
third and that tone 2 had two positions, either a semitone or a whole tone
beneath 1. This flattening was indicated by a special accidental. When,
however, these innovations took place, the position of the tones 4-6 was
also altered, but assuredly not in the way of the tradition of which Burnell
has been a witness.
In my own vocal performance of samans I am wont to sound the dronetone on my violin in those passges which were probably sung by one
priest; and the nidhanas I sing with greaterloudness; however I have not
yet made a detailed stU9.y about the division of the chant. In the staffnotation I only use the lines and leave the interspaces open as an expression
for my uncertainty about my surmises, for in this way no associations
will arise with Occidental music and the reader's imagination is left free.
After the pitch ofthe tones the musical metrics will have to be described.
In general the musical metre rests on the poetical metre. There is, however,
a difference. The versification, namely, obeys a similar rule as in Greek
and Latin, thus in Sanskrit prosody the theoreticians distinguish between
heavy and light syllabies, guru and laghu. A syllable is light when it
contains a short vowel followed by one or no consonant; vowel-length
or any following consonant-group causes heaviness. In the metre of chant
there is only a difference between long and short vowels, the character
and number of consonants between two vowels being irrelevant. Syllables
with a short vowel are reckoned to la-st one matra or musical time-unit,
syllables with a long vowel last two matras. Long vowels with an r or 0
(zero) over them are sustained beyond their normal duration. The vowel
of a syllable at the end of a parvan or melody-member, lasts three matras,
and in case it is followed by vikrti-tones, the prakrti-tone keeps the length
of three matras 1). Sometimes for sacro-musical reasons a short vowel of
the text is lengthened etc. At all events the samans are never to be divided
in bars based on the numbers 2 and 3 and multiples which only contain 2
and 3 as factors, as is the rule in European classical music. The chant is
divided in parvans, melody-members, marked in the notation by a dal).la,
and in the performance by a pause, at the beginning and end. Sometimes
a parvan contains a complete pada, verse-member, or even more than
one pda; generally, however, the pada contains several, a greater or
smaller number of parvans. Since the metre of the chant is based on the
metre of the verse, the writing and printing of smans should be based
on the verseform, with stobha and text characterized by different lettertype.
As a minor no te to this part of the introduction I hav to remark that
in the printed staff-notation I do not indicate the length of the tones and
in my vocal illustration I give this length only approximately, since all
the details in this respect are not yet satisfactorily clear to me.
1)
Cf. Simon, p. 525 s.v. svara 2.
15
After pitch and duration the spiritual contents of the chant should be
characterized in brief. In these ancient saci'al songs the ancient Hindus
address themselves to their beneficial gods, yet this world is besides
crowded with a host of harmful beings. So the chant is a means not only
for pleasing gods but also for imposing upon devils. For this reason and
other magical purposes the udgtar sings his part of the chant in the
most mysterious forms; in one case for instance he replaces all the syllables
by the vowel 0, or he retains, a very commoI'l expedient, only the vowels
of the syllables and he substitutes bh for aU the consonants and consonant-groups; and in general the text of the chanters is subjected to a
great many changes, so is often prolonged to -i, 0 to o-i, or the
vowels and diphthongs i, i, e and ai are replaced by -i or yi, a diphthong of which the components are clearly divided in the articulation,
although in the synable-counting of the melody it is reckoned as one
syllable. Remarkable is the way in which a pda is cut up into parvans
without taking into account the semantic connection of the syllabies.
FinaUy stobhas are added, either meaningless but very holy syllabies, or
expressions of wish, quite loose from the context. All the time the chant
remains destined for the omniscient gods, who see through all these
mutilations easily, and very likely read immediately in the soul of the
chanters. And therefore these chandogas ought to remain conscious of
the thoughts and wishes wrapped in their mysterious chants. And the
modern investigator should consider himself to be one of those gods to
whom the verses are dedicated and he should take all precautions to look
through the magical adornments. And one of the first means to this
purport is to sing the chants to himself and to copy the smans in verseform with distinction of pu!?pas, stobhas and text and with notes as to
the translation and ritual use.
At this point I have finished the general part of my reading and I am
now co ming to the exposition of details. Here I wish to consider the
relation between the spoken and the chanted Vedic language. Leaving
aside the accentuation of contracted syllabies, i.e. the theory of the so
called independent svarita, we can say the following: the spoken Vedic
language moved over three pitches. In the first place this rule was valid
for the daily spoken language of the Vedic age, but also for the softly
repeated prayers of memorization, the japas, on the sacrificial field. On
the other hand the reciting priests were allowed either to stick to one and
the same pitch, in other words to speak in a strict monotone, or to use
an undulating accentuation which according to sacral rules differed in
certain respects from the tone-movement of the bh!?.
The researches of Herman Oldenberg have taught us that we may take
PQ.ini's description of the ac cents asa basis and that the ways of accentuation followed by the priests in their sacrificial actions may be considered
as artificial forms. There is, however, in my opinion, one exception in the
16
Smavedic recital, which seems to be due to dialectical influence and wiIl

afterwards require our attention.
But now, let me first start with an example and choose for this purpose
the hymn with which the ~g- Veda-sal!lhit opens. This hymn, as a
matter of fact, is likewise found in the Sma-Veda, but because of the
sacrificial context in which it is used, we find the text in the ral}.yakaSal!lhit and its melody inthe Aral}.ye-geya-gna, the text- and chantbooks which the chandoga-pupils were aUowed to study exclusively in the
wood, far from human surrounding.
The text (1;tV. 1, 1, 1) runs:
agnim ile purohita1Jt
yajnasya devam rtvijaml
hotara1Jt ratnadhatamam.
This text we shall first accentuate accorling to the rules of Pl}.ini.

But before doing so I want to make a remark. In the introduction to
PJ).ini's A~tdhyyi the great grammarian Patafijali warns his pupils
not to pronounce in speech the vowels as if they were chanting. This rule
may have been useful for pupil and contemporary; now whilst showing
the often occurring dependence of melody on spoken language, I shall sin
in two respects, I shall speak in a rather chanting way, and although
leaving the relative position untouched standardize and measure out the
intervals between the syllabIes; on the other hand I shall sing in a speaking
way. I acknowledge both mannerisms are incorrect, but they are preferabIe
to the way in which German clas~ical and romantic poets have imitated,
or so they think, Indian and Greek verse. My intonation gives at least
some idea of how ancient Greek and Sanskrit may actually have sounded.
For, whereas we, Occidental nations, give more attention to one syllable
of a word than to the others by means of more breath and loudness, the
ancient Hindus and Greeks attained the division of attention over the
phonal side of the word by tonal movement and a clear distinction of the
syllable-length. The strophe, then, which I have read to you in Occidental
school-fashion and in which the poet compares the God of Fire successively to a house-priest, a priest of a great sacrifice and a bounteous
sacerdotal r:eciter, runs in Pl}.ini's bh~ as follows:
J;l.V. 1, 1.
Bh.
I.
1
~im
s
s
s s
, , , ,
1
l'
[tviiam I hotdrarp
ife P'!!rohitarp Y!!iiiaaya devam
-
s I
l'
ratnadhtamam
All the words here except the second have one syllable bearing an udtta, so agnim is oxytonon, etc., but the word ile as verb in the main
clause is enclitic. Once we know the place of the high accents the further
voice-melody is known. For whenever the rule can he applied, tone 1 or
17
STUDIES ON THE s.MAVEDA
udtta is prepared by a low tone 3 (sannatara) or at the beginning of a

clause by a series of such tones and it is followed by a curved aftertone
or svarita. This regular succession for instance is found in the fust four
syllables agnimile and in the second four syllables purohitam, but in the
following four syllables there is no room for the middlepitch tone 2 as
the vowel of de- has to prepare an udtta. And between devam and rtvijam
there is no room even for the curved after-tone as the preparation of an
udtta always preponderates.
From the bh~-accentuation we pass over to the Smavedic accentu
ation, that is the accentuation used in memorizing the text. It runs as
follows:
SV. (S. 3, 4)
81281231282812
purohita~ yaiiiasya evam rtvijam
agnim f4e
12
hotlJ,ra~
812
ratnadhtamam
Here we have to do in the fust place with the general Smavedic rule
of orthography that a syllable without any digit of accent has the pitch
of the last syllable which had such a digit. Therefore the second syllable
of ile is pronounced on accent 2 and in the third pda the tone of the
syllable -fii,- is carried on over -ra."" and rat-.
As to the ac cents themselves, when we compare the Smavedic with the
bh~-version we notice that the udttas have remained: -gnim, -ro- etc.,
with one exception: the syllable -vam of devam. Similarly the preparatory
low tone remains low: ag-, pu- etc.; likewise the middle.pitch tones are
unaltered: -g,e (-Ie), -ta."" etc. The most important change is the change
of the curved aftertone into a middle-pitch tone as we perceive in the
syllables i-, -hi- etc. This is probably a sacerdotal prescript. Of another
nature is the lowering of the syllable -vam of devam; here we meet with an
intricate accent-rule running: tone 1 is replaced by tone 2 if there is no
opportunity for tone 1 to flow out into tone 2 in the following syllable; a
consequence of this rule is i.a. that tone 1 on the last syllable of a pda
closing a phonetical unity is always replaced by tone 2. Even a series of
tones 1 is changed into a series of tones 2 when the last of these tones
lacks the required possibility of flowing down. This accent-rule is so in
tricate that we had better &scribe it to an automaticunreflected law of
sound that to a conscious prescript. In this connexion I have to mention
a striking peculiarity: whenever the melodization of a. text is strongly
bound to its accentuation, in such cases tone 1 of the general bh~ (and
likewise of the J;tgvedic recital) is maintained. This tule rather agrees with
Caland's hypothesis that originally the smavedic sacerdota.l pupils had
no rcika of their own, but used the J;tgveda-saIphlta, or that an anthology
from it was prepared for their needs. So the accent-rule, just formulated,
is of a relatively recent date. Now I take up the J;tgvedic a.ccentuation,
IS
which I des cri be according to the theoretical textbooks and the modern
grammars based on them:
~V.
~im
S.
*
".
.S. . s
"
"
"
P"!:rohita7Jt Yc:iiiasya d~vam rtvijam
"
"
hotra7Jt ratnadhtamam
Here wc see the high tones of the bh~ lowcred to middle pitch,
although they are still called udttas, a fact which has two consequences:
Uy, the curved aftertones rise above all the other syllabies, 21y, the
syllables which were formerly udtta are on the same pitch with the
pracaya-syllables. Consequently the syllable -gnim is on the same pitch
as the second syllable of the wordirf.e. The J;tgvedic accentuation is evidently based on a sacerdotal prescript. Likewise the freedom, given to the
hotrakas to recite their hymn in strict monotone, is of sacral nature; it is
a peculiarity reminding us of the bh~ which demands monotony in
speaking to a person at some distance. Not only for the recital but also
for the chanted verse th is solemn monotony is of great importance.
From the accent-systems of the spoken language we now pass on to the
Smavedic melodization. Again I first take an example: the strophe which
I have now examined three times in succession. In its Smavedic performance with its vocalic changes (pu~pas) and interpolated ornaments
(stobhas) it runs as follows: 1)
AGG. 4,1,13. I./mano hu/vayo hu/varco hu/=/22.2./22.2./22.2./
BI. 11 287. II./agnim/irf,i/purohitrn/ = /21. /112' /323 234 5. /
I1I./iha hufi<.J hu/yur Uu/ = /22.2. /22.2./22.2. /
IV./yajiia-/sya di-/vam rtvijm/= /21/1.1 2/3.23 234 5./
V./s\lar hu/jyotir hu/rtarp hu/= /21. 1./22.2./22.2./
VI.lhot-lra'f!t r-Itnadhtmml = I III 1.123/ 32 345 5656'/
VII. le/mahbl = 123 ./ 13 2345' I
Thc thl <le lines of the verse-strop he are here increased to seven lines
by the addition of the stobha-lines I, lIl, V and VII. These stobhas
consist partly of words, partly of meaningless but very holy soundsyllables such as hu and the vowel e. The words used seem in th is case
to have a meaning: manal}" vaya~" varcal}" mind, vital power and appearance
indicate man; iha, irf., yul}" 'here', the beverage of strength and the
duration of life express the earthly life of the priest; svar, jyotil}" rtam,
heaven, heavenly light and divine order indicate the godly in this world,
and maha!}" extension, encloses the three spheres indicated.
The lines I1, IV and VI, which together contain thc text, give rise t,1)
the following remarks. Only in line 11, the first text-line, the division
into parvans agrees with the grouping of the syllables as parts of words,
1) For the methods adopted for the transcript ion of accentuation, melodization
etc. see infra Chapter I Section 11 (Historical Evidence), Preliminary Notes.
STUDIES ON THE SAl\iAVEDA
l!)
but this is merely fortuitous, the line of 8 syllables being dichotomically

divided into two groups of 4 and 4 syllables and the first of those groups
again into 2 and 2. The lines IV and VI follow this scheme of division
again automatically. The first parvan of line II agnim is in accordancf'
with the bh~-accentuation; we meet with a similar case in the first
parvan of line IV, where the syllable -jiia- bears the udtta; however,
in both cases the preparr..tory low tone is replaced by a middle-high tone.
In line VI the syllable ho- has kept its udtta, but in the following syllable
the originally curved aftertone has been replaced by tone 1, a caSt' of
often occurring 'acoustic metalepsis'. The dichotomy applied in the three
lines is a much used means in melodization.
The second parvans of these three lines are completely detached from
the bh~-accentuation; they begin on pitch 1, which extends over the
beginning of the following syllable; th en there takes place a des cent in
line II and IV to 2 and in line VI over 2 to 3. In all these parvans we
meet with vowel-Iengthenings or diphtongizations: rji, di, r. The third
parvans of the lines II and IV are built on the same pattern, the third
parvan of line VI forms again a climax to this.
Line VII consists of two stobhas, the syllable hal), at the end is lengthened and on its vowel the chanter sings in slow time the tones 32:: l.-):
such a descent, in th is case first deviating towards 2, but always fi.nally
reaching tone 5 is called a svrya and tone 6 which lies beneath this tOIll'series is called the atisvrya.
In general the three verse-lines des cri bed are independent of the bh~
accentuation with the exception of their first parvan. In the stobha-lines
I, III and V sacral monotony predominates, but svar contracted out of
suar with falling movement in the bh~, has got an ascending form, with
acoustic metalepsis of tone lover hu.
My further exposition I shall divide according to the following plan:
Ily, stobha-clauses which, placed in a sman, keep the melodic movement
of the bh~-accentuation except in their ornamental closing; 21y, a
stobha-line and a sman following in their melody the bh~-accentu
ation, however with exception that the curved aftertone is mostly replaced
by tone 1 ; moreover in a few cases even the preparatory low tone bet ween
two udttas is raised to tone 1; 31y, two smans for which sacral monotony
is characteristic; th is monotony is sometimes strict so that a whole pda
is sung on tone 1 or tone 2, but very often the last syllabie is ornamented,
and likewise the first syllable of a parvan sung for the rest on tone 1, is
lowered to 2. A fourth form of melodization, which here, ho wever, will
not be discus8ed, is the breaking up of the pdas into three- or foursyllabic parvans moving in their prakrti-tones on the pitches 1 and 2,
and sometimes 3.
GROUP 1. Here I give two stobhas as examples. In the first the

singer l'raises his own song which MS been directed to God lndra, and
20
thinking of his own song he caUs it asmabhya1{t gtuvittamam, 'most

wholesome to us'. In the bh~-accentuation it runs:
Stobha SV. 1. 1322
Bh.
I.
1"
~hya"",
1 "
giU~vittamam
Sung at the end of the sman it has the following melodic form:
SA.
I- - .
/
--.
212
~hya"",
2lt I
r128
gtuvitta~2345m
Essentially the melodic movement has been left unaltered, but as is

always done in such cases, the curved aftertone and the preparatory low
tone are replaced by the middle pitch; the last vowel is lengthened and
provided with a svrya that deviates at the beginning. Tone 3 is consequently only touched in the svrya.
The second example is of a similar kind. In the sman itself the chanter
praises the pressed out, golden soma-juice that haa taken its seat in the
leap of the sacred fire and the added stobha-clause expresses the wish:
may we overcome all dangers. In the bh~-accentuation and its chanted
form the sto bha runs:
Stobha SV. 1, 51 Pl.
Bh.
I"S S I.
.~
..
1"
"-
...!2!ZJ
ati vivni durita tarema
SA.
I~
1 2
2r
Ir
2 r
/ ati vivni durit tarem234 /
The melodization follows the same ruIes as the stobha of SV. 1, 1321,
given above; however the descending gamut ends with tone 4, an interruPted svrya.
GROUP 11: cases in which the melody follows the bh.-accentuation
except in the syllables with acoustic metalepsis of tone 1.
The fust example is a stobha-clause. The sman to which it helongs
is a magical chant on the occasion of a fire. The poet acknowledges that
the God of Fire and the fire-causing God of Lightening have come to
earth according to legitimate disposal ; earthly witnesses of authority can
affirm this. H, then, in general the poet acknowledges the right of these
gods and is grateful to them, in this special case he entreats them to he
lenient. The focus of the fire is called in this oonnexion the womb towards
which the evoked Gods have gone.
21

Stobha SV. 1, 9()1.
Bh.
I
1
"-
yon!m indraA ca gacchatool),
Sii..
...---8
6"
1r
/ yo234nim / yonim indraA ca gacchathO,23451), /
In the bha""aform -nim as preparation is low, -draS has the curved aftertone, the conjunction ca and the verb are enclitic. In the melody the fust
parvan contains the word yonim with great melodie movement and the
second parvan consists of the complete stobha, in which both the preparatory low tone of -nim and the curved tone of -draS have become
high tones.
The saman which serves as second example has for contents an exalted
praise of Indra; if the God should possess a hundred earths and a hundred
heavens, his own splendour and extent would surpass the effulgence of
their thousand suns and their wide extent. The verse belongs to the
strophetype which normally possesses 8 + 8 + 12 + 8 syllabIes, a frequent
form of pragatha; 8yulp and 8uryalp, however, are contracted and remain
so in the chant-form. The samavedic memorizing-accentuation will first
be given. When here a syllable has received tone 2 instead of tone 1
because this tone would not get the opportunity of flowing into a following
tone 2, this has been indicated by italic type.
SV. 2, 212. ya(d) dyava indra te satarp,
BI. III 431. satarp, bhumr uta 8yulp
na tva vajrint 8ahasrarp, 8urya anu
na jatam ~ta rodas
1.12.22.2.32.
31.12.32.2.
1.2.22.323.23.23.
2.31.23.122.
The accentuation here takes place according to the samavedic methods,

all curved aftertones are replaced by tone 2 and the preparatory low tones
have relOained. The rule for lowering 1 into 2 under the prescribed conditions I have already mentioned, in the melodization the l's are here
restored.
The chanted saman runs as follows: 1)
AGG. 1,1,3. /yad dyava indra te satam/e/
/satarp, bhumr uta/8yova/
hy.G. 1, 1,7. /diBa:rp. visa:rp. hasj
BI. V 387.
/na tva vajrint 8ahasrarp, 8urya anu/
/asva sisumati/
/na jatam ~ta rodas/it/iQ.a/
BI. Il 388.
=
=
=
=
=
/1.11.22.2.21./1/
/21.11. 21. /11/
/11.11.1./
/1.1. 22.212.12.12. /
/11.1122./
/1.11.12.112./1/112345/
Here according to the rule found by Oldenberg all the curved aftertones of the bha~ have been replaced by 1; but moreover in the second
1)
I follow here the reading of the Ohyagii.na with Oldenberg's emendation.
22
syllable of the word asva of a stobha and in the first syllable of the word
jatam in the last text-line the preparatory low tone between two udttas
has been altered.
GROUP III, two instanees in which the monotone predominates. The
first example is a well-known hymn to Savitar in gyatri-form (J;tV 3,
62,10-12; SV. BI.V p. 601). It is unnecessary to give here the accentuation of the text; in chanting-form it runs:
ltat savitar vare1Jiyoml

11.111.11111
Ibhiirgo devasya dhimiihl 122.222.222.1
Idhiyo yo nal}, praco-I
12 1. 1.1.12'12'121
Ihuml l dayo/l
11/12111/23451
Here the first pda forms one parvan sung on 1; the second pda one
parvan sung on 2; the third pda is divided into six syllables forming one
parvan, and two syllables hidden in the closing stobha-line. In the parvan
of six syllables the first syllable is put on tone 2, then tone 1 is continued
over four syllabies, the last syllable is ornamented; the two avagrahas
here indicate very short pauses or, as one could say, an accented beginning
of a new tone.
As second example of the monotone I have chosen a hymn belonging
to the hyagna samvatsara-parvan : 2, 1, 3, BI. V 421, AGG. 3, 2, 12.
Text: S. 2,7 BI. II 272 sqq., specially 274, J;tV. 8,89 (78),5. I follow
the version given in Caland and Henry 256a.
/yj jyj j-/yath apilrviy auho/hiyA./auho/ = /1. 21. 2/11.1111. 232/2211/232/
/mghavan vrtrahatyy(j
" / " / " / = /111 . 11111 .232
12211/232/
/Uit prthivim aprthay
" 1 " 1 " / = 11.111.1111.232
12211/23 2/
ltd astabhn uto divm
" / " 1 " 1= 11.111. 11.11.232 /221 1/ 232/
Ivh uvl e/ payl:J.I = 112.223 /2 3/ 1123(s/l)
Here pda I is divided into 2 + 6 syllabies. These first two syllables

are by an unaccurate doubling made into four syllables which undulate on
the tones 1 and 2, and the six following text-syllables bear monotonously
tone 1, the stobha-addition undulates on I, 2 and 3. The two parvans consisting exclusively of stobhas likewise undulate on I, 2 and 3. The three
following pdas are kept undivided and have no tone-movement before
the stobhas. The final parvan-line of the strophe ends in the words payal},
of which the last lengthened syllable has a direct svrya. As to the contents the hymn praises Indra as victo,r over the demon Vrtra and as
creator of the world; and the stobha payal}, expresses the sacrificer's
wordly de sire : milk, that is: a large herd of milking-cattle.
And now I enter upon the last part of my introduction. In this part I
shall, as I said, refer to the grammarian PJ)ni and I shall have to mention
1) The reading of the melody of parvan /ooM uoo/ in the last stobha-line, given
by C. and H" undoubtedly contains amistake, since the syllable -ho of the
preceding parvan has the pralqti-tone 2 and for the second syllable of our parvan
again 2 is given without any other digit between.
23
the adhvaryu. I have to do this simultaneously. For the literary monument

which I am going to describe, is of a very peculiar character. An old
authority such as the Jaiminiya-Brhmal}a does not know what to call it,
either a chant or a recital, and the ritualistic stras allot the performance
facultatively to a special chanter, the subrahmal}ya, or to the adhvaryu,
who as a rule recites prayer,a. And Pl}ini and after him Patafijali describe
the piece as spoken language and the l'ule of accentuation which Pl}ini
gives for its recital is the same as the rule, given by Oldenberg for the
acoustic metalepsis of tone 1 rather independently of his very ancient and
authentic predecessor. The rule runs, as you remem ber: tone 1 is extended
over the following syllable provided that this syllable is not followed by
a syllable which likewise bears tone 1. Moreover Pl}ni here mentions
one exception to this rule, namely when in the line of convocation gods
and priests are invited to come to the day of libations; the acoustic metalepsis which is correctly forbidden on the second syllable of dev!l- is not
applied to the second syllable of brahmr.w!l- either, evidently on account
of a sort of melodic assimilation.
Because of its ritual use and its crude popular contents and careless
versification the piece is also interesting. This subrahmal}y, as it is
called, is sung on the three upasad-days which precede the suty-day of
the ag~t;oma. The performing priest invites God Indra to be present at
the libations, for, as the God will remember, he has committed many a
naughty trick on earth: when the priest Medhtithi had bought somagrass in view of an approaching sacrificial festival, Indra assuming in
hasty gluttony the appearance of a ram ate all the stalks in which the
juice was contained; and when the brahmin Gautama for a short while
was away from home, Indra in Gautama's shape enjoyed the loveliness
of Ahaly; nay even, in order to .make acquaintance with the pleasures
of love on the female side, he became the wife of Vr~anasva, a man so
strong that he put bulls instead of horses before hil> cart. The most important part of the piece consists of an irregular tri~tubh, each line of which
has the caesura after the fifth or sixth syllable; each word-group thus
originating contains either a vocative-group, or a single vocative with
an imperative preceded by a prefix. Now in the bh!tl-accentuation in a
vocative, in a vocative-group at the beginning of a clause, and in a prefix
with imperative the first syllable bears an udtta and all the other syllables are unaccented. But here the two first syllables of such a group
are accentuated. So the piece runs as follo~s:
aubrahmatlyo"m - aubrahmanyo"m - aubrahmanyo"m, =
indra, gaccha, - hariva, gaccha,
medhtither 1n6{Ia, - ""anavasya mene,
gaurvaakandinn, - ahalyyi j(jra,
kauika brhma~, - gautama bruootul,
tryahe suty(jm gaccha maghavan,
dev brahm(J,~ gacchaM, i!9acchata, gacchata.
/2221./2221./2221./
/11.112./112.112./
/1122.22./11222.22./
/11222./1122.22./
/112.222./112.222./
/21.21.112.222./
/12.122.1122.1122.1122./
24
STUDIES ON THE sMAVEDA
Friedrich Nietzsche in admiration of Schopenhauer asserted that the

tragedy was bom out of the spirit of music. Now we face the question
how was music itself bom, and how did music become the expression of
spirit so that we can speak of the spirit of music. Here then, thanks to
Oldenberg, we are witnesses of music's birth and we feel happy at India's
rich literary inheritance which will in the course of time teach us music's
spiritual development. As to the fust point we have seen that the tonal
accentuation, once characteristic for all European languages and preserved
in old Greek and Sanskrit, offered its contribution to this happy event no
less than folk-dance and the discovery of musical instruments.
As the following chapters will be more technical, I should like to formulate here my impression of ancient Hindu music. Any one penetrating
into Vedic chant will be struck by the fact that this chant certainly gave
little opportunity for personal inspiration, but that all the same this chantsystem in its totality rests on religious emotion. It is in this respect as
with the fashion of dress in a remote district; it does not give any
chance to private choice, but in its innate nature it remains an expression
of aesthetic sentiment.
CHAPTERS I AND 11,

in which the author defends more fully the prepositions laid down in the Introduction and in
which he classifies and analyses some groups
of smans as a tentative answer to the question
of how to study the smavedic gnas.
BIBLIOGRAPHICAL TABLE
In the chapters and footnotes the books of this table are referred to by the
name of the author.
Benfey, Th., Die Hymnen des Smaveda, Leipzig 1848.
Helmholtz, H., Die Lehre der Tonempfindungen, Braunschweig 1863.
Grassmann, H., Rigveda bersetzt etc., Leipzig 1876 and '77. (An easy introduction
for students of ancient music).
Burnell, A. C., The Arsheya brhmatta, the Sanskrit text edited together with ....
an introduction .... Mangalore 1876.
Burnell, A. C., The Sa1{&hitopanishadbrhmatta, the Sanskrit text with a commentary
etc. Bangalore 1877.
Oldenberg, H., Die Hymnen des Rgveda, Bd. I, Metrische und textgeschichtliche
Prolegomena. Berlin 1888.
.
Ser;;agiri Sstri and Railgcii,rya, A descriptive catalogue of the Sanskrit Manuscripts
in the Govemment Oriental Manuscript.Library, Madras, vol. I Vedic Literature.
Madras 1901-1905.
Caland, W. et V. Henry, L'agn!loma, Paris 1906.
Simon, R., Das P~pastra mit Einleitung und Uebersetzung. Abh. K. Bayer.
Akademie der Wiss. I. Klasse XIII. Bd. lIl. Abt. Mnchen 1908. (Abbreviation:
PS).
Felber, E., Die ind'ische Musik der vedischen und der klassischen Zeit. SitzBerichte
der Kais. Akademie der Wiss. in Wien. Phil. Hist. Klasse, 170. Bd. 7. Abt.
Wien 1912.
Clements, E., lntroduction to the study of lndian music, London 1913.
Fox Strangways, A. H., The music of Hindostan. Oxford 1914.
Macdonell, A. A., A Vedic Grammar for students, Oxford 1916. (On thewholeihereader
will find here sufficient information. For a few details works such as Wackernagel's Grammar and Benfey's Introduction in his publicat,ion Die Hymnen
des Samaveda (1848) are required).
Hoogt, J. M. van der, The Vedic Ohant studied in its textual and melodic form.,
Wageningen 1929.
Caland, W., Paiicavi.""sa-Brhmatta, Calcutta 1931.
Dresden, S., Algemeene Muziekleer. Naar den llden druk van J. Worp-S. van
Millingen geheel opnieuw bewerkt. Groningen 1931.
Alain Danilou, Northem lndian Music, London and Calcutta, 1949. (The book,
which scarcely bears on Vedic music, is not referred to in the following pages).
The Sii.ma-Veda-SarphitA is quoted from the edition of Satyavrata Sii.maSrami,
Bibliotheca Indica, Calcutta, 5 volumes 1874-'78. Abbreviations: SV. 1 = PA. =
Prvrcika; SV. 2 = UA. = U ttarrcika.; AS. = r8J.lyaka-Sarphit; GG. = Grii.megeyagii.na; AG. = AraQyegeyagii.na; "OhG. = "OhagAna.; "OhyG. = "Ohyagii.na; BI. =
Bibliotheca Indica; I - V refer to the volumes.
CHAPTER I
THE VEDIC AND CLASSICAL HINDU MUSIC WITH
SPECIAL REFERENCE TO THE TON AL SYSTEM
PRELIMINARY NOTE
Besides the many fanciful renderings of the old smans given by modern
smagas, many of which can be found in Fox Strangways' chapter on
the Sman, we possess two traditions delivered by sacerdotal schools and
noted down by the trustworthy witnesses Burnell and Se!jlagiri Sstri.
The latter has not very clearly worded his description, which fact I shall
afterwards return to.
According to Burnell the smavedic gamut runs as follows:
(7)
According to Se!jlagiri Sstri it should be read:
(7)
Both authors agree in accepting that the numeral 7 is an indication of

a note which would be the lowest of the gamut but that in this meaning
it is never found in the Smavedic gamut. Further they both Ie ave out
the kru!jlta, tone 11, higher than tone 1 and occurring only twice in the
grmegeyagna.
In both traditions the minor third is taken as the distance between the
tones 1 and 3, and a diminished fifth between the drone-tone 5 and one
of the predominant tones on which the chanters sing their solos; according
to Burnell this diminished fifth is between 1 and 5 (f-b) and according
to Sef:lagiri Sstri between 2 and 5 (eB--a).
In my Introduction I gave as my opinion that these intervals of a minor
third and a diminished fifth cannot reach back to the Vedic period. It
sooms more likely to me that the smans were originally sung in a pentatonic scale either as ede/ag or as ag/ede. Yet when we meet nowadays in
the Sma-veda-saIphit with a tone 2-flat indicated by 2 we may conclude
that once at a time the interval 1-3 has become a minor third so that
Burnell's interpretation of 1-3 as led would correspond to 2-natural and
Sef:l8ogiri Sstri's rendering as I(es)d to 2-flat.
28
STUDIES ON THE sAMAVEDA
The exposition in this chapter will be arranged according to the following plan:
Section I. The general line in the development of musie. The discovery that musical intervals as psychical facts are based on vibrational
fractions as physical causes. The modern Hindu sruti-system. The septimal
tuning due to foreign influence. . . . . . . . . . . . . . . . p. 28
Section Il. Historical evidence. The smavedic rcikas and gnas.

The numerals 1-6 and the indication 2. The BrhmaI).as and the terms
upagtar and sthna. The PUlllPastra and its commentators, the notion
udha; the relation between accentuation and melodization in the PS.
R. Simon's term 'type' rendered by Fox Strangways as 'melodic figure',
character of these types. The Nrada-Sikl? . . . . . . . . . . p. 35
Section lIl. Modern tradition. Bumell and Se!;lagiri Sstri. Fox
Strangways' account of Bumell's and Se!;lagiri Sstri's investigations.
Se!;lagiri Sstri's description of the J;tgvedic recital. A note of Bumell on
the mutual relation of tones in the gamut. Historically untrustworthy
sman-renderings noted down by Felber and Fox Strangways . . . p. 47
SECTION I
THE GENERAL LINE IN THE DEVELOPMENT OF MUSIe
In the study of the Vedic chant a great deal of work has to be devoted
to the theoretical treatises translated by R. Simon. From the PancavidhaStra and partly from the PU!;lpastra we leam the division of the smans
into solos and nidhanas. As soon as we have actually leamt this it will
be of help in the study to characterize the solos by the added dronal tone.
Reasoning thus it occurred to me how unfit Bumell's rendering of the
smavedic gamut is with b-f as distance between dronal tone and tone 1.
And then I remembered what I had leamt in my schoolboy-days from
Helmholtz' most impressive book,the first edition of which I have put
down in my bibliographical tabie. According to authorities quoted by
Heimholtz many folksongs, specially Chinese and Keltic songs, have
preserved the pentatonic scale; Helmholtz took up by way of explanation
the tones fcgad, whereas (in my Introduction) I preferred the tones of
our stringed instruments, but this is immaterial. From this importance
of the pentatonic scale follows that mankind has standardized the tones
of its voice by musiCl instruments. The fifth and its supplement in the
octave, the fourth, were thus the first tones fixed and so there arose the
gamut cfgc'; then second fifths from f downwards and from g upwards
were added: (bes)cfgc' and cfgc'd'. Having got so far by fictional reasoning
I found the following affirmation in a book on the General Theory of
Musie by S. Dresden (1931). From this work I quote with a few omissions
the following passages:
29
"Our present day gamut of seven tones has had in the course of time
many predecessors of fewer notes. Probably that of three notes, for instance elg, would have been the oldest, then followed series of four tones
(tetrachords) in different grouping of the tones. . .. By putting two
tetrachords together, either next to each other or so that the last note
of the first tetrachord was the first note of the second tetrachord there
originated a gamut of seven tones. In the Greek musical system the
tetrachord!! were represented as descending, by which order the connexion between the different steps and the basis or final tone was closely
expressed. "
"But before these arrangements of tetrachords into one gamut there
must have existed a gamut of five tones, a pentachord,which was in use
not only in Greece, but among many Oriental and Occidental nations,
for insta nee the series edega. For convenienee' sake and by way of comparison we could describe this pentachord as a major gamut with omission
of the fourth and seventh notes. "
"In this old gamut of five tones there have been composed, even in
later days, several themes and motives, i.a. Beethoven : Overture Leonore
no. 3 Allegro; Grieg: Peer Gynt Suite no. 1 'Morgenstimmung'; Mahler:
Lied von der Erde, the main motive."
"Finally we have to mention the hexachord." - Under this heading
Dresden ment.ions the three hexachords of Guido of Arezzo ( 1000 A.D.):
edelga, gabede, Iga(bes)ed.
To this passage I should like to make the following annotations. The
note common to two adjacent tetrachords would be called in Sanskrit
the aIpsa, 'the tone which divides' (1), on tl\e function of this aIpsa in
Hindu music Fox Strangways has an interesting passa,ge; according to
him tone 3 was the &Ipsa of the original vedic gamut. This seems very
likely when we accept the pentachord as original, but in the Pu~pasiitra
the tones 4-6 are parallel with the tones 1-3.
To the Hindu mind as weil as to the Greek mind the pentachord ran
in descending line, although in tuning the upper tones were based on the
lower tones as we know from the expression udiiha. The descending
direction is clearly shown in the final svryas: 345, 2345 or 32345, and
12345.
As to Dresden's remark that even modern composers have used the

pentatonic scales there arises the question of how to explain this tendency.
It would be a superficial answer to say that these composers heard pentatonie gamuts in folksongs and imitated them. For in what way did these
folksongs have acharm for them 1 I should prefer saying with the psychoanalysts that mankind has an emotional memory passed on through the
generations. Similarly dadaism which we find in the chants of the Smaveda and in' folksongs has revived during these latter decennia in the
literature of Europe, and in the same way the quintal pentatonic scales
of antiquity maintain their capacity of expressing emotiollS.
30
STUDIES ON THE Sl\IA VEDA
The discovery of physical vibrational fractions as the

basis of psychical musical intervals. - The discovery of mathematical perspective is one of the great facts in the development of Occidental pictural art; this discovery coincides with the beginning of modern
European history. A fact of similar importance for music was thc arithmetical study of inter vals ; the beginning of this study took place in the
days of Pythagoras and his school. Undoubtedly Pythagoras was a very
great man, philosopher, geometrist and musicological mathematician,
although it may be uncertain whether he was a great discoverer or th at
his merits should be attributed to thc fact th at he no longer shared the
popular Greek prejudice of foreigners being barharians and that he learnt
difficult foreign languages not for the sake of commercial profit, but in
view of spiritual wisdom. At all events just as in Europe mathematical
discoveries promoted the development of music, so a similar fact separates
in India vedic and classical musie. And any one trying to understand the
smavedic gamut and the classical sruti-system, must acquire some
elementary notions of acoustics.
As su eh elementary notions of acoustics I sum up: I. a musical tone is
due to vibrations arising in the source of sound, either the human organ
of speech or a musical instrument, and it is transmitted to the human ear
through a medium; the Hindus ascribed this transmissioll to the kSa,
a term which we could translate by 'physical space' . Arist, ,tie taught
th at air transmits sound and in the heginning of modern history this
theory was accepted in consequence of t.he discovery of the air-pump. 2. A musical interval is based on the arithmetical proportion between the
number of vibrations of two tones. Supposing one tone to have 800
vibrations per second and the other 1600, th en their vibrational fraction
is and their interval is caUed an octave. Similarly a pure fifth is based
on a proportion of ~ and a pure fourth on ~.
The theorem which I mentioned in the second place, has two important
consequences:
2 a. When two intervals are added, then, physically their vibrational
fractions are multiplied, and substraction of intervals means division of
fractions. For instance, when to the interval c -- g = ~ is added the interval
g - d' = ~, then there originates the none c - d' with the fraction ~ x ~ = ~,
and when from this none is substracted the octave c - c', ~hen there
remains the quintal whole tone c' - d' with the fraction ~ : =~. And
when we add the fifth and fourth together, then we get an octave, ~ x ~ =
2 b. By means of logarithmical calculation it is possible to express
physical vibrational fractions in intervals. In European practice the
gamut has been divided into 12 semitones of exactly the same width, i.e.
with the same vibl'ational fraction. And the theoreticians of music h~,Ye
gone farther on this road and for the purport cf calculation ttey divid7
the octave into 1200 cents. If we now call the vibrational fraction of a
cent g, then 1200 cents added together will give an octave, thus q multi-
31
STUDIES ON THE S:\IA VEDA
plied 12PO times with itself will give thc vibrational fraction of an octave
= f. Consequently q1200 = ; and 1200 log q = log 2, thus log q = ~~.
If we then wish to express any interval = rt in cents we get the equation
.], = q", thus log a = x log q or
_ log a _ I
. log 2 _ I
1200
- I
x - log q - og a, nOO - og a X 0,3010300 - og a X 3986.
Examples: lst example: how many cents are contained in a pure
fifth? Fifth = a = ~, thus number of cents = x o,~ log ~ x 3986 = log 1,5 x
x 3986 = 0,17609 x 3986 = 702 cents. - 2d example: which is the
number of cents contained in a whole tone? (Quintal) whole tone = ~;
x = log ~ x 3986 = 204 cents.
Whereas Greek music alld generally speaking, European musie was
supported by the theoretical insight in the nature of intervals, at all
events in Hindu music a minute direct taxation of intervals must have
separated Vedic and classical music. In order to get an insight in this
development it is desirabie to take one more step in the European theory
of music.
3. The intervals are divided into harmonie or restful and disharmonie
intervals. The harmonie intervals are characterized by fractions of which
numerator and denominator are low digits, specially 1 - 6. \Ve get a
series of gradually smaller harmonie intervals wh en we take those of
which the vibrational fractions are: the octave, ~ the fifth, ~ the fourth,
~ the (melodious) major third, ~ the (melodious) minor third. Fifth and
The g.string of the violin and its flageolet-tones.
r-
1/5
T"
gil d" bol
d"'b" i
~
~
~
~
~5
fourth are each other's supplement in the octave, ~ ;: ~ =

major and
minor third are complements to each other in the jjfth. ~ x ~ =~. The
octave minus the major third gives the minor sixth, : ~ = ~. Here the
denominator is 8, and the interval is harmonie. IntPf\'als with 7 or a
multiple of 7 in numerator or denominator will be'is('ussed afterwards.
The intervals mentioned here have the following numbcr of cents, the ,
32
octave 1200, the fifth 702, the fourth 498, the major third 386, the minor
third 316. All these harmonious intervals in their correct melodious form
the violinist can produce on his instrument without any inaccuracy,
specially on the two lowest strings; let us take for example the g-string.
For thls end the violinist plays the flageolet-tones, putting very lightly
the nearest fingertop against the point which divides the string in the
proportion of ~ and ~, of i and ~, of { and ~, of ~ and ~; then he produces in
pure melodious tuning the tones g', dil, gil and bil, all flageolet-tones which
are in practical use. Moreover by putting hls hand in the first position
he can after some searching divide the string in the proportions ~ and ~,
Th& flageolet.tones on the g.string.
g'
dil
gil
bil
dil'
and ~, and so produce bil (the same as mentioned before) and dil'. By
different combinations he can thus produce the intervals mentioned above.
As a rule, however, the violinist plays th tones by pressing the fingertops firmly on the strings and he gives the intervals according to hls
European training in tempered tuning, e.g. the major third with 400 cents
instead of 386 cents required for the melodious major third.
Disharmonie intervals and intervals resembling the melodious intervals
can be easily found by addition or abstraction. Here I give only a few
examples. The major whole tone, Le. the whole tone arrived at by quintal
tuning, for instanee: (c-g) + (g-d') - (d-d') = c-d; ~ x ~ : = ~;
702 + 702 - 1200 = 204 cents. The minor whole tone, i.e. the whole tone
arrived at by the combination of quintal and tertial tuning, for instanee :
(c-e) - (c--d) = (d-e); ~: ~ =~; 386 - 204 = 182 cents. The difference
of these two kinds of whole tones = ~ : ~ = ~, or 204 -182 = 22 cents.
More examples will be met with in the passage on the Hindu sruti-system.
Now we turn to the intervals with a vibrational fraction which contains
7 or a multiple of 7 in numerator or denominator. For their explanation
we consider the horn. The hornist namely can play a long serial of partial
tones by his embouchure, i.e. the attitude and tension of hls lips whilst
blowing. The tone, however, whlch theoretically corresponds to the total
volume of air in his instrument, he cannot produce. Now, this series, if
we take the C-horn, is the following.
Partial tone:
(1)
2:
Fraction:
Cents:
:l
J
f
1200
702
r 6Ol
11
J J pr
'98
888
3111
10
rL~
10
t
2117
281
20' 1811
etc.
33
Here we see that the hornist can easily produce the major tone of 204
cents and the minor tone of 182, whereas the one whole tone used in modern
European music amounts to 200 cents. Moreover he can produce a small
minor third above the sixth partial tone. This minor third g'-bes' (septimal) has a number of 267 cents against 300 cents of the minor third in
equal tempering and against 316 cents of the tertial minor third. When
we substract from the septimal minor seventh (i or 969 cents), for instanee
c'-bes', the melodious major sixth (~or 884 cents), then we get the 'septimal
semitone' of 969 - 884 = 85 cents. Peculiar enough the Hindu srutisystem uses this septimal semitone twice in its gamut, but it does not
use the septimal seventh, neither does it use the melodious major sixth
in the most common gamut. Evidently the kind of semitone here described is due to Arabian influence.
The Hindus have revealed great genius in their construction of musical
melody; here we only occupy ourselves with their tonal system. By
listening sharply they discovered, like the Greeks, that the major whole
tone is wider than the minor whole tone, and they found the tertial semitone by distracting the major third from the fourth; that is according to
our calculation ~ : ~ = , or 498 - 386 = 112 cents. Comparing the intervals: major whole tone, minor whole tone and tertial semitone they took
their proportion to be 4 : 3 : 2, and so they spoke of 4, 3 and 2 srutis,
taking the sruti to be smaller than the tertial semitone. Now these three
tones have 204, 182 and 112 as number of cents, that is, simplified, the
approximate proportion of 9 : 8 : 5. So the subjective taxation of the
Hindus was far from correct, but it gave them a means for naming the
intervals, and the rest was done in singing by auditive correction. And
with the help of their musical intuition they built up a great many gamuts
of which they described the function which the several steps in them had
got: final tone, drone-tone etc. The tones used in the different local gamuts
were afterwards collected in one series, the so called Indian sruti-system
which runs as follows:
1. septimal semitone = septimal seventh minus melodious
major sixth
85 cents
112
2. tertial semitone = fourth minus melodious major third
"
182
3. minor whole tone = ~ =
"
204
4. major whole tone = ~ =
"
294
5. artificial minor third = tertial semitone plus minor tone
"
316
6. melodious minor third
~ =
"
I;
386
7. melodious major third
,
"
408
8. quintal major third = 2 x major tone
"
498
9. fourth = !S =
"
HJ. the artificial fourth as quintal major third plus tertial
520
semitone
"
590
fifth minus tertial semitone
11. the enlarged fourth
"
SlO
12. the diminished fifth
fourth plus tertial semitone
"
34
3
702 cents.
fifth
2
787
the septimal sixth = fifth plus septimal semitone
"
814
the melodious minor sixth = 1200 minus 386 =
"
884
the melodious major sixth = 1200 minus 316 =
"
the quintal major sixth = fifth plus quintal major
tone = 1200 minus 294
906
"
996
18. the quintal minor seventh = 1200 minus 204 =
"
19. the tertial minor seventh = 1200 minus 182 =
1018
"
1088
20. the tertial major sevellth = 1200 minus 112 =
"
21. the artificial major seventh = melodious minor sixth
1108
plus artificial minor third
"
1200
22. the octave = f =
"
When we substract the inter vals which follow one another we get as:
differences 27, 70, 22, 90, 22, 70 etc., which were all called srutis! Yet the
melody-formation of the Hindus has its merits and impresses other
nations, and as to its tone-system it found its origin in the discovery of
major and minor whole tones. Seeing now how late this discovery was
made in other countries, we may surmise that the smavedic gamut was
originally of a pentatonic character. Only whell the width of the accents.
of language altered and the tonal base of music got another aspect, the
distance between the numerals 1-3 could be changed into a minor third
and the digit 2 could get a double meaning (led and f(e8)d), but then.
simultaneously the term 'udha' i.e. harmonic derivative in upwards.
direction got both the meanings of fifth and fourth, so that 12345 forinstance meant fedca with different intervals between f-c and e-a.
Before leaving this subject it may be useful to consider the so-called
sa-grma in relation with the sruti-series. This sa-grma consists of the
note~ ~ja or ~ (sa) = c; fl:!abha or ri = d, 204 cents; gndhra or ga = e,.
386; madhyama komal or ma komal = f, 498; paiicama or pa = g, 702;.
dhaivata or dha = a, 906; nil?ii.da or ni = b, 1088; ~a4ja = c, 1200 cents.
Two things here deserve our attention. In the first place the names.
sa and pa have only one meaning, (very rarely at least a flattening of ~
takes place); they indicate the tones c and g; all the other names have
by means of accidentals four meanings. The notes c and gare simply
called suddha, the notes ri, ga, dha and ni are called in their sagrmafunction tivra, on the other hand the tone ma. is in the sagrma called ma.
komal, whereas ma tivra is fis of 590 cents.
In the second place the intervals ~a-pa, ri-dha, and ga-ni are all equal
to 702 cents. These two facts taken together seem to indicate that the
sagrma has developed out of a hexachord without the fourth. This hexachord had thus the form cdegab, which in its turn arose out of cdega or degab;
in other words the development of the tonal system in India a.nd Europe.
followed similar lines in the beginning.
13.
14.
15.
16.
17.
35
SECTION II
HISTORICAL EVIDENCE
THREE PRELIMINARY NOTES
NOTE I. The indications of the changes in a text when

melodized.
1. As a rule such changes are left here to the reader to find out by
the use of the Smaveda-edition. However, when expressed, the following
principles will be applied :
2. . The melodization of textwords will only be given by melodynumerals the words being omitted, but the sound-stobhas are given b9th
in the articulate and melody-form with the mark = between.
3. The 'split' diphthongs, i.e. the long diphthongs ai and au when
distinctly articulated as two sounds but prosodically possessing the
value of one syllable are spelled as ai and au, but never as ayi or avu.
Analogically the split diphthong i will never be printed as oyi.
NOTE 1I. Method adopted for transcribing the accentuat.on.
In order to avoid the difficulty of printing numerals above letters
the following rules of transcription are adopted:
1. Firstly the text is giveu, then the numerals of the accentuation
with the mark = bet ween ; full stops indica.te the end of words, sometimes the semicolon indicates a metrical division.
2. When a tone is prolonged over a following syllable, the digit is
repeated.
3. The letter r indicating that a dependent svarita follows more than
one udtta is omitted.
4. The numeral 2 when indicating that in Smavedic accentuation
an udtta. (or a series of udttas) is lowered when not, (or when not
finally) , followed by a dependent svarita, is printed in italic type.
5. The u which follows the first of two or more such 2's in italic type
is omitted. (Vide e.g. SV. 1, 20 pda. c).
6. The mark I is used to drawattention to the fact that the preceding
pda c10ses a phonetical unity, cf. Macdonell, p. 438 n. 2, and p. 449a.
NOTE lIl. Method adopted for transcribing the melodization.
1. In order to avoid the printing of digits above the worM, prakrtitones are indicated by large type, e.g. 1, 2, and vikrti-tones by small type,
e.g. l' Z
2. In general the lengthening of vowels will not he indicated; if it is
strictly necessary, it will he done in an added text-line. The letter r indicating an extra. lengthening of a long vowel, will be omitted.
3. When a. tone is prolonged over a following syllable, the numeral is
repeated.
36
4. Whcn the mark follows the digit 2 in the same syllable, this is
spelled as 22; when this mark is placed on a syllable following a syllable
marked by this digit, this is expressed as 22; when a vikrti-tone bears
this mark, it is written as 2.
5. The numeral 7 is replaced by 2_1, cf. Simon Einleitung p. 517 s.v.
abhigita; the avagraha when indicating a vinata, is replaced by L2'
cf. Simon. Einl. p. 522 s.v. vinata.
A
DOCUMENTS OF HISTORICAL RESEARCH
The docurnents of historical evidence belong to different periods. We

may divide them as saIp.hits, brhmaJ.las, siitras and sikss.
SA~HITAS. The Smaveda-sarp.hit uses several numerals as
indications of toneR and their duration. For the interpretation of these
numerals, we have, as far as the saIp.hits are concerned, the important
fact that the numerals 1-3 are used in the Arcika of the Smaveda for
the indication of accents, numeral 1 indicating the udtta, numeral 3 the
sannatara and numeral 2 not only the original pracaya-syllable, but also
the svarita which in the bh~ deseends from 1 to 2 according to PJ.lini's
rules (1, 2, 31 sq.). And when thus 1-3 form a descending series, it is
likely that the same thing will hold good for aU the nllmerals 1-6.
Now there is a peculiar indication in the Smaveda-edition (BI.) which
desen-es specialattention, namely the combinations 2'2 and 22; thcyare
as a rule, as appears on a simple perusal preceded by tone 1 or 2, and
followed by 3. For further illustration we shall consider some of the
'types' (parvan-melodies) mentioned by Simon in his Introduction to the
Pu~pa-Siitra (p. 511 sqq.). The commentators of this PS. give onIy the
nume:als ofthe vikrti-tones leaving it to the reader to look up the prakrtitones in the Smaveda itself.
ndhlJ, = 122. The meaning of this type is defined by SV. 1,3132 The
verse-strophe consists of a tri~tubh, of which all the pdas follow the same
melodization. Each pda forms a parvan, and the two last syllables are
everytime repeated twice with lengthening of their vowels. The same
couple of stobha-parvans follow these repetitions, whilst a stobha-line
precooes and follows the chant-strophe; at the end two new stobhaparvans are met with. The chant may be symbolized as follows, full
stops indicating the end of words:
/ihi = 123/ ihi = 122/ ehiy = 3322/ ov = 3234 5/ hi = 2/
Ist pd:/ 222.22.22223.123/ 12~/ 32:,/
/
/
4th pda:/ 22.2.22.222.23123/122./ 3322/
/
/
/ihi = 123/ ihi = 122/ ehiy = 3322/
/ hi = 234/
/auhova = 553/ i = 31345/
So we soo that what in Simon's type ndhlJ, sooms to he a sym.bol 2z at
the end, really precooes a following 3. Of the applications of this type
37
1 only consider here PS. 8, ll2, which prescribes an alteration in SV.

2, 902 sq., sung on the Naudhasa-tune, BI. V, 63. The Nau<lhasa-tune
itself meant here is given SV. 1, 2365 , BI. I 487; in this yoni the first five
syllables of the third line abhi vatsa'Y(/, na are prosodically altered into
bhi vatsa'Y(/, na and melodized as /1232/11.2/. Now the first pda of SV.
2, 903, which as belonging to the second verse-strophe of a pragtha,
becomes the third line of the second chant-strophe runs in the rcika:
accha hi tva; the vowel a is lengthened and i diphthongized and these
syllables form two parvans /iiccha/ hai tva/ = /1232/12/. The alteration,
then, comes to this: the type /ll2/ is changed into /12/ with lengthening
of the vowels, whereas, however, no notice is taken of the flat 2.
tuvis~ma?t = /32235/. The meaning of this type is defined by SV. 1,
457. Strange is the fact that here 22 is placed on a short vowel. Verse 1,
457 consists of 4 x (12 + 4) syllabies; the word tuvis~ma?t is there used
as the four syllables which form the closing parvan of pda a, see BI.
I 909. In PS. 8, 147 this type is applied in the description of the saman
Lausdya to the text SV. 2, 502-504, BI. IV 136. The first four syllables
of pda a in the first strophe run according to the arcika: pro ayiisid; now,
in the yoni (SV. 1,557; BI. II 180), these four syllables forming one parvan
are melodized as /23 2345/, the augment a being alided after pro. The
varavantam-type /2(2)3 2345/, defined by SV. 1,173, BI. I 121 is here in
the iiha, where the augment is restored, changed into /3235/, vide BI.
IV 136 line 5. Here again the flat is ignored.
naimis cakrauva = /ll.122/. The type is defined by SV. 1, 94. The
words quoted run in pda d of thercika: nemis cakram ivbhuvat; in the
chant they are altered by pu~pas and stobhas into:
[naimis cakrauv/iva/ bhuvat/ = /ll.122/ 3 234 5/ 21/.
Here again in the sman 2 is found before tone 3. The type is applied
by PS. 8, ll9 to the second and third chant-strophe of SV. 2, 463-465,
a sman of which the text is given in SV. 1,446,445 and 444. In the siitra
mentioned it is said that only in the first of these three chant-strophes,
thus in 2, 463 the nemis-type is left unaltered. Now in the arcika 2, 463
runs as follows:
pra va indraya vrtraha'Y(/,tamaya
vipraya gatha'Y(/, gayata ya'Y(/, jujo~ate
The five last syllables of verse-pda. d (vide BI. IV 100 1. 10) are by
and stobhas altered into
pu~pas
/ya'Y(/, jujau-va/up[~ato Mi/
/1.12.23/2/1'22'3'52/
While then in SV. 1, 94 naimis cakrauva is melodized as [11.122/, in

SV. 2, 463 we have ya'Y(/, jujauv with one syllable less, melodized as
/1.1223/' thus the three tones 1 have become two tones 1, and a still
greater licence has been taken by changing 22 into 223.
38
miidalJ, = /22323,5/, where 2 is onee preeeded by tone 3. The type is

defined by SV. 1, 5785 , a sapha-sman, where this parvan is based on the
word madalJ, at the end of pda b of the text. In the sapha-r.man SV. 2,
446, BI. IV 74, referred to in PS. 10, 83 the type is extended over three
syllables /ir/iiniim/ = /32 2323,5/. Here tone 3 is prolonged from the last
syllable of the preeeding parvan and further the flat is omitted above the
2 oeeurring between two tones 3.
Our final eonelusion may be that the textual tradition about 2 is very
uneertain, and further that tone 2 is seldom preeeded ,by 3 but as a rule
by 1 or 2. From an alternation as 2and 23 it is likely that the mark ..
indieated a flattening. Should the flatiening of 2 amount to a semitone,
then it is likely that the distanee between 1 and 3 had beeome a minor
third. At that time, whieh I do not reekon to be original-Vedie, 123 could
have the value of led and I(es)d; the form e always oecurred when a long
passage, for instance a parvan, was sung against the drone-tone 5; es was
only used as a transitory tone, its express 'indication was
At all events, and this fact is worth emphasizing, there is disagreement
between the definition of the udha as given in the Pu~pa-Stra and the
tone 2-flat with its eonsequenCeB. And I am inelined to explain the difficulty in this way that I take the udha-definition as bearing on the
original pentatonic scale of the Smaveda, whilst I attribute the tome 2flat to an introduction of later da.te, when the tonal system of the Smaveda had changed of character. I do not think that we have proof conelusive that this alteration had taken place before or during the redaction
of the Pu~pa-Stra, as the complete examples are only given in the commentaries.
A.
BRAH~AS. - We turn from the SaIphits, in this case the SmaVeda-SaIphit, to the brhmaJ}.as. This class of literature is of interest for
us with reference to the terms wpagiitar and stkna, this word were to
mean saptaka.
As early as the year 1855 the term upagtar was registered by the
Petropolitan Dictionary as found in three brhmaJ}.as and brhmaJ}.a-like
texts: the Taittiriya-SaIphit, the Aitareya-BrhmaJ}.a and the Satapatha-BrhmaJ}.a. Caland and Henry (p. 173) give the following note:
"Bebind the ehanters [at the main chants of the a~~ma, to begin
with the bahil;t-pavamna-stotra] the subchanters (upagtra:Q,) take
their places, at least three in number. They accompany the chant singing
constantlyon low tone (mandra-svareJ}.a) the syllable ho, only being
silent when the chanters sing together, in the nidhanas. The sacrifieer
himself likewise joins in as a subchanter but on the syllable om". - We
may conclude that the usa of tone 5 as a pedal tone is of old date and
goes back to Vedie times.
The term oaptaka is not given in the Petropolitan Dietionary (volume
VII anno 1875) as a Inusicologica.l term. It is mentioned by Clements in
39
the meaning of octave. Clements (p. 16) writes: "The Indian system of
vocal music aIlots three saptakas or octaves to the voice, each saptak
ranging from s up to ni." In the staff-notation on p. 17 the mandra
saptak ranges from F to e, the three saptaks together from F to e". "The
soprano voice would be an octave higher according to Indian ideas."
The informationis not quite clear, in Europe we are accustomed to the
idea that one individual voice does not extend even to two octaves.
Now, in the Paiicav~a-BrhmaJ}.a 7,1,7 bearing on the chanting ofthe
gyatra, it is said of the chanter: mandram ivgra iidadittha tiirataram
atha tratamarp, tad ebhyo lokebhy'o 'gsit.
Calandtranslates this: "He should begin softly, then [chant] louder
and then still more loudly; thereby he chants in view of these worlds".
And he comments upon it in the foIlowing words: "mandram, trataram
and tratamam could equaIly weIl mean 'deep, higher, highest (pitch of
the voice) as relating to the three octaves (grmas, sthnas) which are
equaIly designated as mandra, madhyama, tra (Taittiriya-prti. 22, 11)
and said to reside successively in the chest, the throat and the head (ib.
10. and cp. Nradiya-Sik~ 1, 7: ura?t ka'TJtha?t sira caiva sthiinni tri'TJi
vilmaye), each of these sthnas comprising seven tones."
Now, one can understand that the ritualist in view of the three worlds
extending one above the other prescribes that the chanter in bis three
repetitions of the songs should use a higher pitch every time. But that
these distances should he an octave apart in the European sense of the
word, it is difficult to understand, and I doubt whether by the references
to the Taittiriya-Prtikhya and the Nradiya-Sik~ we may accept
that the Vedic chanter in the oldest times had already found a division of
the octave such as from C to c into 7 steps. In tbis connexion it is useful
to bear in mind that the use of seven tones by the Kauthumas is extremely
rare and that most Smavedic chanters used a much more limited gamut
(Simon p. 524).
STRAS AND COMMENTARIES. - From the BrhmaJ}.as we pass
on to the PUfllpa-Siitra and its commentators. Richard Simon in bis
edition of tbis work refers to originaIly two and ultimately three commentaries : 1. a PUfllpabh~ya by Ajtaatru (i.a. in Ms. F.), 2. a Phulladipa
by DikflIita RmakrfllJ}.a, who is also caIled Nn Bhi and who is a son
of Tripthin Dmodara (in Ms. N.), 3. a Phulla-vivaraJ}.a by Ajtaatru
(in Ms. T.), Simon p. 485 sq. On p. 499 Simon summarizes the results of
bis analysis of the PUfllPasiitra, and what he assumes to he a proof for
the relative chronology of the composing parts; in tbis connexion he
mentions that the commentary of Ms. T. contains a quotation from
Halyudha, a lexicographer ofthe tenth century A.D., DikflIita RmakrfllJ}.&
is again later than Ajtaatru.
The commentary of Ms. T. ascrihes the PUfllPasiitra to VararuciKtyyana, the grammarian dated ootween PJ).ini and Pataiijali, and
40
the commentary of Ms. F. ascribes the PUIllPastra to Gobhila, an author

of a Grhya-Stra belonging to the Sma-Veda.
With reference to the date of the PUlllpastra the opinion of Caland
about the relative chronology with regard to the Smavedic gnas is of
importance. In his introduction to the PancavIpsa-BrhmaI).a (p. XI) w~
read: "It can be proved with certainty that these two gnas [the haand byagnas] belong to the later strata of the Smavedic literature:
that they are later than the grme- and araI).ye-geyagnas, later than the
PancavIp.sa-BrhmaI).a, later than the ~eya- and KllIudra-kalpa, later
than the stra of L~yyana-DrhyyaI).a, later even than the PUlllpaStra" and p. XVII: "If a chanter knew by heart hls two gnas [grmeand araI).ye-geyagnas] and if he knew which verses ought to be adhibited
in any rite, he could bring about all the changes that were necessary for
adapting a melody to a given triplet or pragtha. These rules for adaptation
were then fixed and systematically arranged in a special book: the PUlilpastra. But, in order to have at hand for immediate use the smans so
as they were to be adapted according to the rules of the PUIllPastra, two
main books were composed 1. the hagna and 2. the harahasyagna
(called by abbreviation the hyagna)."
Moreover, as we have seen, more than ten centuries, if not fifteen
centuries separate the commentators from the original texts. In those
centuries the accentuation-systems of Sanskrit h&.d completely changed,
it had become a language with emphatic accent, which means that in the
days of these commentators the chandogas 'sang' their rcikas as weU
as their gnas. Moreover in their quotations of smans these commentators use the numerals only for the vikrti-tones; the pupilhas to find the
numerals for the prakrtitones in his manuscript of the Sma-VedaSaIp!tit or he has to remember them from oral instruction.
Two passages in the PUlllpa-Stra are of special interest to us with
reference to the fixation of intervals. In the first place 8, 90: caturthamn,ndriitisviiryii1}iim svarii1}ii1ft dvyantaram uccam uccam uduhalJ,. The
udha is the tone which is two intermediate tones higher than the fourth
tone, the mandra and the atisvrya. From this definition foUows that the
tones 4-6 follow the same order as the tones 1-3, a fact ignored by the
Nradiya-Siklll-passage, quoted by Burnell and Se~giri Sstri. The
etymological meaning of udha must have been 'a harmonic derivate in
upward~ direction' ; this allows both the intervals of a fifth and a fourth.
The second passage in the PUlllpa-Stra is 9, 30:
iircika1ft nidhana1ft nyiiye
staubhikarp, vii yad ak~aram/
kr~!iikr~!arp, bhavet sviiryam
a1Ltodtta1J~ '/:rdhc.9varam/ /
In tbc yoni (= nyye) the nidhana is formed out of rcika-words or a

stobha. The last syllable belonging to this nidhana whether it is prolonged
41
(Sim on p. 518-520) or not prolonged, "becomes svrya, and when it is

udtta it becomes vrdhe-svara.
Note. In this' interpretation given by Simon antodiittam = 'having
an udtta on the last syllable', and by contraction kr~tiikr~tam = antakr~tiikr~tam, both adjectives referring to nidhanam found in line a.
Simon explains this passage more fully on p. 524; his interpretation
is largely based on the notes of the commentary N. Here four cases are
distinguished:
Case 1. Such a nidhana-passage may in the spoken accentuation end
in a dependent svarita, thus have an udtta on the penultimate syllable,
then in the yoni the last syllable becomes krl?ta and in the iiha it becomes
a svrya stretching from 1 to 5;
Case 2. Such a passage, when spoken, may end in an anudtta syllable,
Le. as follows from comparison with the next case, in a preparatory low
tone as can only occur at the end of a pda which does not close a phonetical unity 1); in this case the last syllable remains akfl?ta in the yoni and
in the iiha it becomes a svrya stretching from the 2d to the 5th tone;
Case 3. Such a passage, when spoken, may end in a pracaya-tone;
the last syllable then remains akfl?ta in the yqni, and in the iiha it becomes
a svrya stretching either from the 2d or 3d tone to the 5th tone;
Case 4. Such a syllable may end in an udtta syllable, this syllable
remains udtta in the yoni, and in the iiha it becomes a vrdhe-svara.
If we put the four cases in tabellic form, we get:
(rcika or stobha in spoken
form):
Case
Case
Case
Case
1
2
3
4
dependent svarita .
anudtta (i.e. sannatara)
pracaya
udtta . . . . . . . . .
yoni:
ha:
ka~I;la
svrya 1-5
svrya 2-5
svrya 2 ot 3-5
Vfdhesvara 321
ak~aI;lB
aka~I;la
udtta
The differences, then, between the formulation of the PUl?pa-Siitra and

the interpretation given of the Siitra with the aid of commentary Nare
the following: 1. in the siitra three cases are distinguished, and by N four
cases by subdivision of the second siitra-case; 2. the commentator clearly
distinguishes three forms: the rcika-accentuation, the yoni-melodization
and the iiha-melodization. The second point containing the addition of
the accentuation is, as we shall see, of great importance.
In his third note to PS. 9, 30 Simon mentions that commentary N gives
as nidhanas formed out of stobha-words: pitii devaniim with karl?aI)a of
the last syllable, and pavasva soma with aka~I)a. Moreover in the fourth
note to the same siitra we meet with the stobha mahiint samudra(l, as an
instance of an udtta in the accentuation of the last syllable. Now, when
1)
MacdoneIl, p. 438 n. 2, and p. 449a.
42
STUDIES ON THESAMAVEDA
we peruse the staubhika chapter of the Sma-Veda (Van der Hoogt

p. 87 -123) none of these phrases are mentioned as stobhas, on the other
hand we are struck by the fact that they are, all three, syllable~pentads
with an iambic cadence; and indeed they form part of SV. 1,429 and
SV. 2,591, the first of a strophe-triplet forming "OhG. 5, 1,9.
With reference to the commentator's notes the melodizations of 1, 4291,
'dhma-sman' and of 1, 4292 'dharma-sman(SV. I 873) and "OhG. 5, 1,9
(SV. IV 247), ' 'dharma-sman' are of importance. The name 'dhmasman' is derived from the last word ofthe text 1,429. Probably dharman
is a mutilation of this 'dhman'; it is, however, used as a stobha in 4292
of which the last two parvans run: /e/dharma/ = /23/112345/; vide BI.
I 873 1. 9. This melodization is an instance of the rule given PS. 8, 190
in the expression p'rvii:n,gabhtam and in the note to it in commentary N,
which will be afterwards d8cussed.
I shall now begin by quoting the rcika SV. 2,591-593 with its
accentuation (cf. :!;tV. 9, 109, 4).
591.fpavasva 80ma mahiint 8amudralJ,f = /122.22; 31.232/

/pita dvana1[t vivbhi dhma/ = /32.323; 231.12/
592. /fukralJ, pavasva devebhyalJ, 8oma/ = /31.222; 312.22/
/dive '[Y!'thivyai sa1[t ca prajbhyalJ,f = /31. 231; 1. 2.312/
593./divo dhartasi fukralJ, piy~aM = /32.312; 32.312/
/8atye vidharman vji pavasva/ = /31.122; 31.222/
Here the fact of the three 'stobhas' quoted by Simon, forming part of
SV. 2,591 strikes the reader immediately. The melodization of SV. 1,
4292, BI. I 873, runs, with the omission of the stobhas, as follows:
/pavasva 8oma/ = /1l2. 22/

/pita devnam/ = /21. L1l1l1L23/
/mahnt 8anudralJ,f = /21.121/

/vi8vbhi dhma/ = /121. llw/
the last text-parvan is followed by the stobha /auhov = 558 5/ etc.

In comparing this melodization with the accentuation we notice (cf. the
Introduction): lly, the dependence of melody on accent, 21y, the smavedic 2 replaced by the original bh~-tone 1, 31y, the preparatory low
tone replaced by the middle tone, 41y, the application of Oldenberg's
rule: metalepsis of tone 1 in the following syllable when this is not again
followed by tone 1. - Only the first syllable of devanam deviates from
this rule by an ornamental melodization. Similar remarks will hold good
when,comparing the melodization of SV. 2,591-593 with the accentuation given above, except (of counre) for the last syllable.
Now, applying PS. 9,30 with its commentary N to the three so-called
stobhas we may lay down the following correspondences:
Comm. case 1, dependent svarita cf. pita devnam 1 ), with ka~Q.a
in SV. 1, 429 2
1) In the strophe itself the syllable -n4m has a sa.nna.tara, but, the phrase being
used as a separate stobha, it must have a dependent svarita.
43
Comm. case 3, pracaya-tone cf. pavasva soma, withaka~aQ.ainSV.l,4292

Comm. case 4, udtta cf. mahnt samudralJ, with udtta in SV. 1, 4292.
So in SV. 1,4292 the rules given above for the 'yoni' are applied. When
we now look up SV. 2, 592 sq. no 2, BI. IV p. 2471. 13 we find the following
melodizations:
Case 1: pita deviiniim = /21. 2'2112346/ (sic legendum 1)
Case 3: pavasva soma = / 112 . 23 2346 /
Case 4: mahiint samudriilJ = /21.132 1/
So we find here the rules given above for the 'Uba' applied.
Thus the rule of the PUlilpa-Siitra is exemplified with this one example
and the cases 1, 3 and 4 of the commentary are confirmed. However it has
become apparent that the relation here between yoni and Uba differs
from the usually accepted definition, cf. Simon p. 510. And as to case 2
of commentary N no instance has yet beeI! givcn, and likewise the svrya
2346 possible for case 3 remains without example.
Let us now first return to PS. 9, 30 and specially to Simon's note 4,
where commentary T is mentioned. This commentary gives SV. 1,5684
as an instance of case 4. This sman with its accentuation runs as follows:
sakhiiya ii ni ~data = 123. 1 . 1 . 222 .
puniiniiya pra giiyata = 2323. 1 . 222 .
iu'f!t na yajilailJ pari bhi}ata riye = 23.2.31.12.222.32.
In the melodization (BI. 11 207 line 4-7) pda a is preceded by the

stobha /ohi = 55/, the pda itself is divided into 5 + 3 syllables = /455.
4.4./558 5/. Here tone 4 corresponds with 1, and tone 5 with 2; so the
melodization is closely based on the accentuation; pda b is divided
against the natural word-connexion into (3 + 2 + 3) syllabIes; repetitions,
pUlilpas and stobhas are applied and the influence ofthe accentuation is no
longer maintained. The same remark holds good for pda c, here the last
five syllables preceded by the sound-stobha e are melodized as:
/e
23/222.321/.
And further, and still keeping to examples to the rules of PS. 9,30 and
commentaries, I want to consider the Asvavrata-sman SV. 2, 1193-95,
'OhyG. 3, 2, 9, BI. V 484. The extensive stobhas of the sman are fully
described in PS. 8, 232. With reference to the text-parvans it is said there:
piidagitis tulyiiwhich phrase praotically means: in all the pdas the melodization is based on theaccentuation. The text of the first strophe with
its accents runs as follows:
abhi viij viBvarUpo janitra'f!t = /32. 32. 3122. 312.
hira1Jyaya'f!t bibhrad atka'f!t BUpa1'1JalJ = 2323.23. 12.232./
riryasya bhnum rtuthii vasiinalJ = /122.31.231.123.
pari svaya'f!t medham rjro jajiina = 12.31.12.31. 222./
44
The melody of these pdas runs:

pda
pda
pda
pda
a: 121. 21. 2112.213 2345 1

b: 12212.12.11.23211
c: 1112.21.121.12323451
d: 111. 21. 21. 21. 123 2345 1 (medham correctly melodized 1)
The last one or two words of the pdas of the three strophes run as
follows:
strophe
pOOa
a
b
c
d
II
a
b
c
janitrm
8Uparru'i~
vasn~
jajn
vi8varupm
Bam babhv
mimn~
a
b
c
avasya ret~
yukt vasn~
yajiio ddhr
bhridv
vipati~
III
last syllables
I
I
a.ccentuation
312
232/
123
222/
3123
2.312/
123
123.12/
31.123
31.222/
2312
312/
case
1
4
2
3
2
1
2
I
2
3
I
1
melodization
213 1345
232 1
123 2345
1231345 (sic
2113 z345
1. 211 z345
1131345
112.11 zau
21.123 235
21.123 2345
1211 2345
213 z345 (sic
COIT.)
COIT.)
Here the pdas I b (case 4), I d (case 3), 11 b (case I), 11 d (case I),
111 b (case 3) and 111 c (case I) comply with N's rules, that is six of the
twelve clausulae. In two pdas we very likely have to do with textual
mistakes: in I aread: 211 23{5' in 111 d read: 211 2345 .
So four clausulae remain (I c, 11 a, 11 c and 111 a), all belonging to
case 2 of N's rules, but having the svrya 3 2345 instead of 2345' So we may
conclude that thc commentary N and Satyavrata Smasrami's edition
in this respect belong to different schools or rather school-divisions, the
latter only allowing 3 2345 where the former teaches 2(32)345'
In addition to th is discus&ion on PS. 9, 30 I want to draw the attention
to PS. 8, 190 and specially to Simon's note 4 teaching that a fin al parvan
containing rcika words and preceded by a monosyllabic sound-syllable
forms a nidhana. The commentary gives here as an instance SV. 2, 101618, BI. V 187; the melodization of this sman is independeilt of the
accentuation; the last words of these three strophes are siikine accent.ed
312, giraJ.t accented 12 and the enclitic no varat; the final nidhanas are
melodized as: Ie = 23/112345/. Above we have met with the stobhaaccumulation leldharmal = 123/1123451 in SV. 1. 4292 ('dharma-sman').
In the second chapter of this book I have ex. ned the smans of the
Grme-geyagna; and there the clausulae often r.emind us of N's rules,
aresuit which is rather surprising, for the Pu~pa-Str8, is considered to
teach the derivation of the Oha-from the yoni's, collected in the Grmegeyagna; yet in an other r~spect my research is neither brought to an
45
end, for the question remains which of these smans are tl'igtar-chants
and which are prastotar-chants, but I like to postpone this question to a
later opportunity when the ritualistic sourees will be likewise adduced.
A minor note in this connexion is finally to be given to the technical
expression vrddhe-svara. From my 'Classification of the smans of the
Grme-geyagna' (I A b fJ) it will appear that the expression vrddhesvara
originally bore upon the oxytonon word vrddhe in SV. 1, 1201 which was
melodized as 22v on the other hand (l.I. I A b y) the oxytonon word
rayim in SV. 1, 221 is melodized as 321 ; so vrddhe-svara must originally
have had a wider meaning than that accepted by Simon p. 523: '32 1 '.
We must not drop the general subject of the PUli'pa-Siitra without
having considered the term 'type' mentioned by Simon p. 511 and translated by Fox Strangways as 'melodie figure' (p. 276), the correct translation being 'parvan-melody'. The latter author writes: "Mention should
also be made of a method of handing down the traditional manner of
singing. Melodie figures which constantly recur are directed to be sung
after a well-known pattem. For instanee, the caderice to the first line of
SV. 1, 1 is
"Other syllabIes, then, which have this particular cadence are said to
be 'done like hi~i". Thus:
i-
8-
bh
yi
yi
yir
"Herr Simon gives about 100 examples, but until the Smaveda notation
has been exactly interpretated it would he misleading to transcribe
them."
Before making any notations on this passage, I want to remind the
reader that the chants of the Smaveda are no music in the modem sense
of the word. The choice of the intervals in which we sing them is therefore of rather secondary importance. We should consider the Vedic chants
as a magical means for preventing the wordE! from heing understood by
the devils. The addition of pUli'pas and stobhas is therefore of the greatest
importanee, and likewise the breaking up of the pdas into small bits,
parvans, without any consideration of the way in which the syllables
belong together as parts of words; the simple melodification is a means
of solemnifying and is characterized by a tendency toward I!acral monotony and the use of a small set of melodie cadences. As to Fox Strangways' tI-eatment of the smans, as a rule he omits the dal}{las and thus
46
neglects the importance of thc parvan. In consequence of this neglect one

gets the impression on reading his exposition as if a type were a kind of
melodic embellishment, but in reality it is the movement of a parvan.
And in order to make thc matter clear, I shall analyse the example he
adduces: the last word of YS. I, 1. This is altered by pu~pas, stobhas and
parvan-divisions into /b auhova/hi{l/; here 1) all the vowels are lengthened
and the r of bar- is dropped before the stobha auhova according to PS.
7, 184. The two parvans are then melodized as /323(555/32344/. Any parvan consisting of two syllables which are lengthened and are sung on
/323(4/ is said to follow the h~i-tyPe.
About the application of this term which is defined by SV. I,}l we
have to consult Simon p. 515, where we find mentioned the sfttras 3, 7;
6,48; 8, 128; 9,74; 10,20 and 10,91. I shall only discuss here the fust
of these sfttras. In PS. 3, 7 we find the rule that in a nidhana-parvan
following the hi~i-type a tlavya (i, , e) remains what it was (at least
qualitatively, so i becomes i). An example of this rule we find in SV.2,25
with the tril,ladhanysya-melody (BI. III (6). The text of this sman
runs:
punnal!- soma dhray
apo vasno ar{lasi
ratnadh yonim rtasya sdasi
utso devo hirar,tyayal!-.
This verse consisting of 8 + 8 + 12 -+- 8 syllables is divided and adorned
with stobhas in the following way, the numerals indicating this time
number of syllabies:
/6 + huhov/2 +
2/2/2/2/
/2/auhov/6/2/2/
/2/auhov/2/2/2/
The text, t.hen, is in general cut up into bisyllabic parvans without any
consideration, of course, for logical coherence; further the two last syllables
of pda a are combined with the two first syllables of pda b. The rule of
the Pu~pa-Sfttra, then, refers to the two last syllables of the second and
third linea of the chant-strophe, here identical with the verse-strophe.
So the syllables -{laai and -dasi are not lengthened here into /-{lsi/ and
/-dsi/, but into /-{lsi/ and /-dsii on the melody-type /323(4/.2)
So the fact of Simon having met with about 100 technical types of
parvan-melody must not lead us to ascribe to the Smavedic chant a too
greM flexibility and wealth of ornament; even the 100 types of Simoil
analysed in this rettpect would confirm this statement. What the ~a1la of
1) In the first parvan of the quotation the BI.-edition gives the wrong
reading VtJ- instead of ba-.
I) The BI.-edition incorrectly gives /1fIW =3 su 5/ and /diJ8I = 3 su 5/. Cf. the
reading which~iri Sstr gives of SV. 1,11.
47
the Greeks were with regard to their later art of sculpture, that the
chandoga-chant is for the later Hindu art of melody.
SIK~S. - On the literary class of 6ik~s to which the NradiyaSik~ belongs, Winternitz; I p. 243, says in g'neral: "many of these
sik~s
are comparatively old and are closely subjoined to some Prtiskhya or other. .. whilst others are of much more recent origin" . Fox
Strangways (p. 259) basing his judgment on Burnell caIls the NradaSik~ quite modern. For us the text is of interest because of the passage
quoted from it by Burnell and Se~giri Sstri.
SECTION 111
MODERN TRADITION
As modern tradition we have to consider the observations made by

Burnell and Se~agiri Sstri on present-day sacerdotal singing. On the
other hand the smans phonographed by ]'elber or noted down by Fox
Strangways are of a rather fanciful nature owing to their choice of performers. However, before dealing with Burnell and Se~agiri Sstri themselves, it may be instructive to see that Fox Strangways is too enthusiastic
as a historian when he tries by too loud praise to strengthen the authority
of these two scholars. This author, then writes (p. 250): "The two versions
of the first hymn of the Smaveda which follow [in F. Str.'s book in
staff-notation] are from A. C. Burnell (r~eyabrhmaJ)a, p. XLV) and
Se~giri Sstri (Descrlptive Catalogue vol. I p. 77); and the notation is in
~ccordance with the direction of the gnas (the Smaveda text as actually
sung) as they understood them, checked in the first case by the verification of a musician, and in the second by the usage as the writer knew
it" .
Burnell himself teIls us (1.1. p. XLI) "It would be useless to give the
complicated notation as used in the S. Indian MSS., and which I have
already mentioned, for these letters amount to several hundreds. The
principle of the modern notation by numbers is far more simpIe. The
seven notes marked by the numerals I, 2, 3, 4, 5, 6, and the last (really
never used) by 7. Of these tbe first is F and the rest E, D, C, B, A, G."
And in a note thEi author adds: "I have ascertained this by means of a
standard pitchpipe. It is also the doctrine of the Nrada-si~ (adhy. 11)
according to oral information" , and then foDows only one sloka. that gives
information on the tones 1-3. At all events Fox Strangways' note on the
verification by a musician represents Burnell's use of a pitch-pipe. And
moroover Fox Strangways forgets to mention that the two smans which
Burnell writes out do not contain the note 6, so the reader feels very
doubtful whether this observer has really ascertained that tone 6 was
rea.lly sung as A. As to ~agiri Sstri, he introduces his rendering of
SV. I, 11 with the worde: "To give a musical idea of the chanting of the
4S
Smaveda, as it ought to be, I give here the first hymn agna yhi in the
Vedic and musical notation". This author does not inform his readers
how he knows that 'it ought to be' as he tells us, yet we may feel sure
that he has heard competent singers sing it like that.
But now there is a very peculiar thing: when we compare the two 'versions' of the hymn mentioned, then we notice - besides Fox Strangways'
omission of one entire parvan and of several dal).<,las - that the two
authors do not only differ in their musicological interpretation of the
numerals, but also that they cannot have had the same reading of the
hymn before them; for with Bumell the hymn ends in tone 4 and with
Se!jlagiri Sstri in tone 5. The first lectio is not only given by the BI.edition, but is moreover confirmed by the hi~i-type of the PU!jlpa-Stra.
Of minor mistakes of Fox Strangways' rendering I mention that in the
penultimate parvan one should read the stobha au-ho-v and in the last
parvan hi{!i and not rhi{!i, the r having been dropped before the stobha
au-ho-v.
At all events from the staff-notation we leam that the notes 1-5
according to Bumell and Se!jlagiri Sstri (when we do not take ccount
of the later remarks of the latter author) should be:
Seiilagiri S~tri (prima facie)
Burnell:
This notation is right with reference to BumelI, but as to Se!jlagiri

Sstri we should reckon with th is author's remark that although he
writes e for tone 2 he really means es; "from the way in which the hymns
of the Sma(veda) are sung it appears clearly that e is flat" (l.I., p. 77).
Now we can further leam from Se!jlagiri Sstri's description on p. 77 sq.
that according to him note 6 is bes. So we find Se!jlagiri Sstri's interpretation to be:
Seiilagiri SSl?tri
(according to bis intention)
On p. 78 namely we read: "The sixth note of the Smaveda, which, as

already explained, corresponds to b flat or kaisiki ni!jlda, and the seventh
which corresponds to g or paicama, occur very rarely 1). So, omitting
these two sounds, I may say th at the melody of the Smaveda is uniform
and may be identified with the rga, called bhogi, which is a derivative
rga referred to the original kharaharapriya, and whose notes are !jIa, ri,
ga, ma, dha, !jIa."
1) In reality 7 in the meaning of tone 9 never occurs, but 6 is found in a
separate class of smans ending in tbe cadence 5151 > and in a few otber instances.
49
Now, whilst I have shown above in Fox Strangways a too great enthusiasm as a historian, I also have to attribute to him a certain carelessness,
which shortcomings (I confess) are quite counterbalanced by great gifts.
All the same I have to quote another passage from Fox Strangways' book
found at p. 263.
Here the author hasfirst explained how according to his ideas the smangamut could be explained by the gndhra-grma - N.B. a gamut based
on the sruti-system! -and then he continues: "There is, however, another
tradition as to the intonation of the sman-scale. M. Se~agiri Sstri gives
the scale as Ma, Ga, Ri, Sa, Dha 1) and says that the sound is that of the
rg bhogi. That rg is:
S
2:
(Numerals added by me, B.F.)
~
Ii
7.l.
(!;
M
Q.
a.
"Drone D (which does not matter here, th drone being a later invention),
a1Jt8a E, omitted notes C and A. This is an old tradition as may be seen
from three passages ... ", h",re follow one reference to an early book of
the Mahbhrata and two references to the Nrada-Sik~. "In all three
passages the five notes are the same, and the last two out of order, this
looks as if the first five were substantive and the last optional, thus:
?:
(SJ
r.7
......
t!2
P N
and this is the scale of Se~agiri Sstri and of the rg bhogi."

To this passage I should like to make the following r~marks:
1. The reference to Mahbhrata and Nrada-Sik~ look rather convincing and the construction of the Vedic gamut as having a transilient
nature, is a priori probable.
2. It is, however, inadmissible to identify the ~Q.ja with d. In consequence of this identification the distance bet ween tone 2 and tone 5_
becomes a diminished fifth, and according to the old tradition of the
Pu~pa-Stra it should be an udha, i.e. a pure fifth or fourth.
3. Too easily the author maintains that the drone-tone is a later
invention, when the fact remains that the upagtars are mentioned in
ancient Brhmal).a-passages.
4. Nowhere else in his book does the author explain the rg bhogi.
5. The scale given here on p. 264 of his book does not agree with the
scale found in the staff-notation of SV. 1,11> given by him on p. 251. But
1)
The gamut when read downwards from 1 (ma) to 5 (dha) is the same 88 the
one given above 88 'intended' by Se~agiri Sstri, only the gamut is now transposed
a whole tone higher.
50
there is a remarkable fact: if the author before writing out this hymn had
read Seljlagiri Sstri carefully he would have seen that by e the tone es was
meant. And then there is agreement, even though there is a transposition.
So it appears that Fox Strangways has taken much trouble to get
information from all sides and from well-informed scholars, but that he.
as a musician, was not enough philologically trained.
Now leaving for the moment Fox Strangwys" we return to the two
forms of information we receive from Burnell and Seljlagiri Sstri
Se~iri Sstri's ga.mut
Burnell's ga.mut
As mentioned before, Burnell gives his information of the year 1876 as

an observation on the chanting of Kauthuma-priests and he thinks that
it is corroborated by a passage of the Nrada-Sikljl which he had then
recieved only orally and incompletely. He gives the full quotation from
the Nrada-Sikljl on p. XIX of his Introduction to the Sa:rphitopaniljladbrhmaI).a (edited anno 1877) without noticing however, that this passage
represents another school-tradition than the one used in the sman he had
listened to the previous year. The slokas of this passage which are
decisive, run:
y~l), s'fIUUJnrp, prathamal),
sa ve,!!-or madhyamal), svaral),f

yo dvitiyal), sa gndhras
trtiyas tv rl}abhal), smrtal),f /
caturthal), I}atf,ja ity hul),
pancamo dhaivato bhavet/
l}aI}tho nil}do vijneyal),
saptamal), pancamal), smrtal),f /
Identifying, then, the notes of the classical ljIalja-gamut with the steps
of the European c-gamut - an identification which is in confesso - we
get the following gamut as Vedic according to this passage:
Gamut of the Nra.da-Si~
ma ga ri sa dha ni (pa)
2: r r r J J J J
1
(7)
That is to say, we get the gamut as it was given by Seljlagiri Sstri.

however, without the two flattenings of e to es and b to bes. The notes 1-3
form a minor third divided downwards into a semitone and a whole tone;
the notes 4-6 do not form a regular descent; 1-4 is a fourth, 2-5 a
fifth, but 3-6 a minor third. Only the two first intervals could be called
udiihas; the information is therefore incongruous with the definition of
the PUljlpa-Stra. The substitution of 4-5 for 1-2, often used in the
prastva of the chants, is a second difficulty.
51
Mter all these discussions I should like to formulate my hypothesis

anew: The Smavedic scale was originally a pentachord either ag/ede or
ede/ag; this is made acceptable by the development of Greek music and
the many folksongs still preserved in this gamut. In case the pentachord
was ede/ag the added note was I, which, however, did not jtttain the
character of a tonic.
When the accentuation-system of the language changed, the musical
interval became perhaps smaller, for at the same time the musical
faculty was more developed and tertial tuning had become a regular
factor of the tonal system. So then the basis of the svrya ledcab, as it is
described in the Nrada-Sikf?" was accepted, which no longer formed a.
straight line. The udiiha-definition of the Pu~pa-Siitra remained as it was,
although contradictory with the musical practice. Mterwards led was in
some cases replaced by I(es)d. The mark 22 was a falling tone to be compared
with the old svarita of the spoken language. The flattening of tone 2 was
inconsistently handed down in tradition. However, 2-flat never occurred
when tone 2 was sustained during !lo long passage or a parvan. The alteration of e into es, and of b into bes, as given by Se~giri Sstri is of recent
date, when the smans were often sung without a drone-tone.
With reference to the smaller intervals, semitone8 and minor thirds,
chosen in postvedic time for the recital, it is of interest to quote Se~agiri
Sstri's description of how the J.tgveda WM in his days recited in hls
neighbourhood (Descriptive Catalogue p. 3 sq.).
"There are three svaras or accents in the Veda: the udtta, 'acute',
the anudtta, 'grave', and the svarita, 'toned'. When more than one
anudtta precedes an udtta, all but the last are often sounded lower. The
first anudtta following an udtta becomes svarita; and all the rest are
sounded like the udtta and called prcaya. If we compare the above
accents with the tones of a musical gamut, we can know the exact sounds
of these accents. Let the udtta be represented by any key on the piano
and be marked as C, and ~ according to the Hindu method, then (1) the
two sounds of the anudtta will he marked as Band ni, and G and pa,
respectively. (2) The svarita will be sounded as C sharp and rio (3) The
pracaya svara, as it is pronounced by the Brahmans of the Taittiriya
Skhya and the J.tgveda, is identicaI with the udtta, and therefore is
equal to C and ~. [Ad 2] a long svarita in the J.tgveda is split into two
long sounds - the fust of which is sounded as udtta and the second as
svarita..
];tv. 1, 1, 1.
ni
88a
84
ri
ri
ni 84
Ie py - 1'0
hi - taf'!l
52
ni
ssa
ri ni
sa ni ssa sa
(j r r#u~~ U
ri
~j:I~
yg,j. na - sy d!!. - vam r-tvi - j'l'Jl
[Second example containing several anudttas at the beginning] _

J;l.V. I, 3, 11.
lJ)
p
CQ -
pa
ni sa ni
di: - yi: - tri
sa
sa
B1l - nr - t
ri
ni
na'l'Jl
Annotations to Se~agiri Sstri's description.

1. After these illustrations Se~agiri Sstri gives a criticism on this
mode of performance. In this criticism, however, the author shows he is
not specially acquainted with the ~gvedic Prtiskhyas.
2. By ri (or ri) the author does not.mean as he says himself ri suddha,
but ri komal; cf. Clements, Appendix A and p. 10.
3. Interesting is the way in which the author indicates the length of
vowel and the prosodie weight of the syllabies. The shortnesb ofthe vowels
is indicated by sa, ri, etc., their length by sa, ri etc.; the prosodie weight
of the syllables is expressed by crotchets and quavers.
4. Of interest likewise, but in another sense, is the way in which the
svarita is treated; in long syllables as c-cis, both long, and in short
syllables as cis short.
5. The most important characteristic of this modern ~gvedic accentuation is that the whole compass remains limited to the tones bc(cis) altogether a whole tone; except when the pda begins with several anudttas,
the compass is extended downwards to (J. I a.m inclined to think that the
wider compass represents an old tradition. Fox Strangways (p. 248)
mentions the compass of a minor or major third for the modern recital
of the Yajurveda.
Mostly fitly in connexion with $e~agiri Sstri's description of the modern
recital may be discussed a remarkable passage in Burnell's
Introduction to the r~eyabrhmaI).a (p. XLII):
"The difficulty in understanding the true nature [of the musical notes]
has arisen out of the attempt to classify the notes, and also to connect
them, phonetically, with the accents. It is not difficult to understand
this by comparison with similar attempts of the mediaeval students of
music. Thus Hugbaud (840-930 A.D.) classified the notes of the plain
chant in the following .way:
~gvedic
Sol, La, Si, Ut

graves
Re, Mi, Fa, Sol

finales
la, si, ut, re

superiores
mi, fa, sol, la

excellentes
53
"In the Indian books on musie there is a somewhat similar elassifieation

of the notes:
udtta
ni!i!da, gndhra
anudtta
fsabha, dhaivata
svarita
!i!a<;lja, madhyama, paieama"
It is unneeessary to dweil on Burnell's prejudiee against the attempts

to conneet the musical notes with the tonal aeeents of the language,
sinee we have seen that this eonneeting of tones with aeeents goes back
to the PU!i!pa-Stra, that is to the latest period of the Smaveda itself.
Neither need we take mueh notiee of Burnell's eomparison of the Hindu
tone-elassifieation with that by Hugbaud, sinee the latter only refers to
the registers of the human voiee. Here I want to express my regret that
no titles are mentioned of those 'Indian books on music' . Still, let us
examine the given elassifieation and the attempted eomparison of notes
with aeeents.
As to the classifieation we notiee that those books bring together
ni!i!da with gndhra: b with e, dhaivata with fsabha: a with ri, paieama
with ~a<;lja: g with e. Indeed those couples of notes belong together as
pure fifths of 702 cents, or to express it in the terminology of the Pu~pa
Stra bag are respeetively the udhas of ede. Likewise !i!a suddha and ma
komal: e-t form a pure fifth. If we then give to the thirds e-e and g-b
tertial tuning, we arrive at the Hindu !i!a<;lja-gamut, as arisen from the
hexaehord ede gab.
As to the relation between tones and aeeents, let us give the following
form to the information reeeived:
'Ia
ri
ga
1na
anud. ud. 8var.
pa dha ni 'Ia
anud. ud. 8var.
We shall then get the following result: the aeeents follow one another
in the order of the ~g-Vedie aeeents, but their total eompass is, as it was
settled for the Yajur-Veda, a minor third. So we may eonelude that the
'Indian books on music' referred to by Burnell are eonneeted with a
1;tgvedie school, a eertain freedom being allowed for the width of the
intervals.
Before closing this seetion I should like to eonsider the performance of
a sman by a modern smaga and eompal'e it with the notation as it is
found in the Smaveda-Sarphit. For this purpose I have ehosen AraI,lyegeyagna 1.2.16 (arka-parvan), BI. II 409. It is eomposed on the text
SV. 2,490 = ~V. 6,7, 1 whieh runs:
mrdhana'Tfl- divo arati'Tfl- Prthivy
vaisvnaram rta jtam agnimf
54

kavi1f1- samrjam atithi1f1- janiJ,niJ,m
san nal1- ptra1f1- janayanta devM /
The verse consists of 4 x 11 syllabies, they are.equally divided and

melodized in parvans of 4, 3 and 4 syllabies; a collection of stobhas is
used as introduction and the same collection but extended by repetition
recurs at the end. The sman then runs:
/hu= 2/hu = 2/hu = 2/
/jyadoham = 2345/jyadoham = 2345/jyadoham
/2111/223 11/2345/
/2111/211/22345/
/2111/2311/2345/
/2111/23 11 / 22345/
a /hu = 2/hu = 2/hu = 2/
/jyadoham = 2345/jyadoham = 2345/
/jyadohu = 2345 5/v = 5/e = 2/
P /jyadoham = 2212/e = 2/jyadoham = 2213 2M5/
2345/
The construction of the melody as far as the verse itself is concerned is

very simpie; the pralqti-tones of the fust parvans are 2111, those of the
second parvans 211, and those of the third parvans 2345. The second
parvans sometimes show a lengthening of tone 2 into 223 or 23 combined
with a prolongation of the vowel. Looking at the total melodization of
the chant-strophe we may characterize it as follows: in the preceding
stobha-part the first three parvans are monotonous on tone 2, and the
last th-ree parvans have a falling type. The four pdas of the verse-strophe
have two parvans staying monotonously on tone 1, but with tone 2 as
indication of the beginning ; the third parvan has the falling type 2345.
The stobhlc conclusion consists of two parts, first a repetition of the
commencement (a), however with the alteration of -ham into -hu and the
addition of /v/e/; the movement of this part a can again be described as
consisting of a monotonous part: hu three times repeated, and a falling
type likewise three times repeated, whilst, however, the third jyadoham
shows a climax with regard to what has preceded; the stobha e on tone 2
is a predecessor of part p. This part p shows for a long time an undulating
movement on the tones 1 and 2, but finally enda in a svrya that deviates
at its beginning.
Now I give the sman as it is sung by a modern smaga and phonographed by Felber. The da:r,:uJas express here short rests, those after the
second parvans of the verse-lines were omitted, I have inserted them as
broken lines. As the phonogram was put in staff-notation by a European
scholar we cannot see in what way the intervals are influenced by the
ruti-system. In hls introduction Fox Strangways (p. 272) La. writes:
"sung by a boy of twelve, a Brahman, in the Sanskrit College, Calcutta".
hu
iyadoham-a
"tE, ~ e"
hlJu
iyadoham-a
55
hu
iyadoham-a
er Iff ij ri r r r
m'r - dha-na~ di
JJI
a - ra - ti~ pr-thiv-y-I),
vd
gJ f er Irrr ir rre:? ~ I
2:l flJ f re If? r ~ i r r r ~ J I
'):1
r-
ooi - Ava-na-. rm-a
kiJ - vi~
8am-r-idm
4 - Ban na1),
pij
ia - tam a- gnJm - a
a - ti - thi~ ia-nIJ - nlJm - a
tr4~
hlJu
I'YIJ
t 4
hu
ia-na - yanta
de, - vd-I),
hlJu
Urrl J ~rrrl~r@rrrlri
4iyadoham-a
diyadoham-a
diyadoham-a
4iyado - hIJ
4iya-do
hlJu - 114
iyadoham -a
ham-a
56
The melodization although to a certain extent based on the traditional

notation has got a totally different character so that historically it is
without any value. As in the present publication I have left rhythm
undiscussed so far, I shall pass by the performance in this respect. As is
immediately apparent all the tonal relations are broken up; for instanee
in the notation the sman begins with the stobha hiiu on tone 2 and
finally ends with tone 5; the performed melody begins with C and ends
with C. Nowhere is the sacral monotny maintained excepto in the three
hiiu-stobhas of the beginning and lass strictly in the hiiu-&tobhas in piece a
of the fiBale. The vowels emphasizing final consonants are not traditional
either. The iijyadoham's of introduction and piece a, which are in the
reading of the BI.-edition exclusively descending, now begin with an
upwards rise. The climax at the end of piece a and in the total piece {3 of
the finale when compared with the introduction is due to the singer's
initiative. As to the melodization of the verse-lines the monotony of the
first parvan has been rtlplaced by a rising movement with an intermediate
note from C to G; the second and third parvans are joined, the former
beginning with an initial rise that is followed by a slowand straight descent
from A to C.
To formulate my comparison briefly: the music has re:rnained Oriental
and Hindu, but it is a sman which the ancient Hindu gods would not
have accepted. The tendency towards monotony has been narrowed
down; the antithesis between monotony and descent has been replaced
by the antithesis between rising and falling, for which purpose the parvandivision is often given up.
Of course the Hindus will go on melodizing their ancient sman-texts
in a modern Hindu way and they themselves and foreign listeners will
most certainly appreciate and enjoy their production. And yet it remains
to be desired that a full historical knowledge about the ancient smansinging will one day be acquired so that a speaker who introduces the
subject may speak about the meaning of the text and the sacrificial
context, about pu~pas, stobhas and parvan-division, whilst an a-capella
choir will represent gtars, upagtars and yajiiapati, and will divide the
chant according to the vidhs and accompany the melody with ho and om
whenever required. In a word, we have to adhere to the postulate of a
musicological philology receiving its enthusiasm from modern production,
but its means and methods from the development of spiritual research.
CHAPTER II
THREE AN ALYTICAL STUDIES ON THE
SMAVEDA-SAMHIT
PRELIMINARY NOTE
The chapter contains :

1. a dassification of the smans of the grmegeyagna according to
their final cadences.
2. an analysis of those smans in the grmegeyagna which end in
the cadence 5656
3. an exposition of Oldenberg's rule with a fuller reference to the
gnas.
A first perusal of the Smaveda teaclies us that there are i.a. the following types of melodization :
a. the melody is closely connected with the accentuation of the spoken
language;
b. it Ehows a tendency to sacral monotony;
c. the text embellished with pu~pas and stobhas is broken up into
many parvans. Very often those parvans contain three or four syllables
which move on 1 and 2 as prakrti-tones;
d. there is a tendency towards contrast in many smans; for instanee
the text-parvans will show monotony or a quiet undulation whilst the
sound-stobhas possess wide movement ; or the opposite line may be
chosen.
Ad a. When the melodization is closely related to the accentuation
of the spoken language the following rules are valid: (cf. Introduction).
1. the preparatory low tone 3 is changed into tone 2; exceptions to
this rule will be found in the following sman-classification in section
I A b E,
2. the change of 1 into 2 in those cases where tone 1 could not flow
out into tone 2 in a following sYllable, as is done in the SV.-rcika, is
ignored a,nd tone . 1 is maintained in the melodization,
3. the dependent svarita is mostly replaced by tone 2 just as in the
Smaveda,-rcika accentuation,
4. often the dependent svarita - but only if used before a pracayasyllable or a phonetical pause - is replaced by tone 1 and thus the tone
of the preceding syllable is continued ('accoustic metalepsis'),
5. according to the general bh~-accentuation a syllable, itself
anudtta:, between two udttas is accentuated as a preparatory low tone
58
(sannatara) and it is so in the Samaveda-arcika accentuation. When the

melodization follows the accentuation, it is generally replaced by tone 2
(cf. rule 1); in a few cases it is replaced by tone l.
The rules for the syllable immediately following an udatta are rather
lax; in the accentuation or melodization of the SubrahmaJ.lya, described
by paJ}.ini (1, 2, 38) dva brhmiirJa dgacchata the syllable -va receives
tone 2 (cf. rule 5 above) and, as is psychologically probable, by the influence
of the word devalJ, (12) brahmiirJalJ, is also melodized as 122 and not as 112.
Cognate to Oldenberg's rule of the change of 12 into 11 is the change
of 21 at. the end of a final parvan into 22 1 , that is the final syllable begins
on tone 2 and ascends to 1, the vikrti-tone. We might call this a partial
accoustic metalepsis, (for examples see Classification I A ba and (J).
Ad c. In case the saman-text is broken up into parvans of three or
four syllables which as a rule move on 1 and 2 as prakrti-tones the following main types can be distinguished:
monotone type
111
222
descending type 112
122
ascending type
221
211
1111
2222
1112
1122
1222
2221
2211
2111
asc.-desc. type
212
desc.-asc. type
121
undulating type
2112
2122
2212
1221
1211
1121
1212
2121
Since in parvans of more than four syllables we often meet with a

beginning like 21111 ... we might also call 211 and 2111 monotonoUB types
with a marked beginning.
SECTION I
CLASSIFICATION OF THE CHANTS OF THE GRMEGEYAGNA
PRELIMINARY NOTES
General remarks.
The classification contains six main divBions according to the pitch
of the final tone.
The following abbreviations are used; a > m means the accentuation
influences the melodization, M = acoustic metalepsis; unacc. = unaccented = without udatta. The chants are counted according to the
nurilber of the Prvarcika and according to the number in the series of
melodies belonging to one and the same verse. Numerals in large type
indicate the prakrti-tones, those in small type the vikrti-tones. As to the
duration of the syllables the reader is referred to the edition. Full stops
indicate the separation between words. In the quotation pw,pas are
often left unnoticed. (Moreover cf. supra p. 35 note In).
I.
59
The diphthong i as a pu~pa for a tlavya is in singing clearly divided

into its two composing vowels and distinguished from the spoken diphthong
ai; for this reason north-indian manuscripts (Simon p. 527) even write
yi; however, in the fixation of the number of syllables in the parvan
this i or yi is reckoned as one, and likewise it has only one prakrti-tone.
As to the division between sentence-stobhas and sound-stobhas I have
in my classification V A a pand V A b b taken iti (3923 ) and vai (467 2) as
sentence-stobhas in accordance with Van der Hoogt p. 92 no. 93 and
p. 93 no. 108 (BI. 11 p. 523 1. 4 and p. 523 1. 6), although vai is clearly a
sound-stobha and iti is taken as such in the classification V B b, for
bisyllabic words as atha, iha etc. are scarcely feIt as words in the smancontext, they have become sacred interjections.
11. Scheme of subdivision, applied in the following subsections.
A. The final parvan consists of words or syllables of the text, or of a
sentence-stobha.
a. the tone is a prakrti-tone.
b. the tone is a vikrti-tone.
B. The parvan consists of or ends in a sound-stobha.
a. the stobha is monosyllabic.
b. the stobha is bisyllabic.
I.
A. The final parvan consists of sentence-syllables or a

sentence-stobha.
a. Tone 1 is a prakrti-tone.
Instances: 151/11.211.221/ 1), 642/21/=rtnn, 1432/121.1/ g~ pad PH
(BI 11 520 1. 5; cf. Benfey, Einl. p. LXIV middle: 'beraus hufig tritt
[in the SV-Pada] in einsylbigen mit udtta versehenen Wrtem kein
Zeichen ein'). - In sman 151 a whole pda forms the parvan : stmarp.
rudrdya drsilcm, M in -mam and -ya. In all the three instances a m.
b. Tone 1 is a vikrti-tone.
a. The parvan consists of a bisyllabic last word of the text: 10 drs,
222 (m), 5682 (-riy), 57()3 (dvitd) , melody in all the instances /221/;
a >m, partial M.
p. The parvan consists of a bisyllabic stobha-word: 12()1 (vrdh), 12()2
(mah) , 175 (vrdh) , 2831 (st~), 3721 (vrdh), 3722 (mah) , 399 (yuddhd,
textword repeated as stobha), 4871 (iha> ih), 5031 (iha ih). Melody
always /221/; a m, partial M.
y. The parvan ends in a bisyllabic oxytonon textword: 221 /agnir no
varp.sate rayim/ = /21. 42. 222.321/, 3932/12.321/' 468'/22.32.1/, 568'/222.
>
>
>
1) BI. I p. 117 1. 2, - here we read the second syllable of 8tomam as ma followed

by an avagraha and a zero-digit. The avagraha probably indica.tes a short pause
in the expiration and the zero an m sustained and sung on the same tone 1 as the
preceding vowe1.
60
.321/,5704/12.32 1/, In all the instanees a> m, and final 3,1> 321, after 2
in last syllable of penultimate word.
15. The parvan ends in a bisyllabic oxytonon word of a sentencestobha: 151 /udadhir nidhiM = /221.12 1/, 190 /dgahiy hi, t im/ =
= /112.11.1.12 1/, 291 2 /mah vis/ = /21.12 1/, 4721 /i~ vrdh/ = /21.121/'
In all cases a > m, M, finall> 21, after 1 in last syllable of penultimate
word.
e. The pda ends in a polysyllabic oxytonon textword; the parvan is
four-syllabic: 4705 (a-)/ghasa'Y(t8aM/ = /22321 /, 506/-bhir asmayuM = /2 .
. 232 1/, 521 (80- )/ma-matsarM = /2232 1/,
C. The parvan ends in a polysyllabic oxytonon stobha-word: 517 8 /viijt
jigivdn/ = /21.12 321/, MP /didihi/ = /232 1/,5544 /viiji jigivdn/ = /21.12 321/
(sic. COIT.). - In all the instanees of e and Ca> m, Min 517 8 and 5544;
peculiar is the maintenance of 3 in the penultimatc syllable or its change
into 23; always finall> 21,
B.
The final parvan consists of a sound-sto bha.

973 /e/ = /2 31 /, 1602 /aihiy aihi/ = /22.32 1/, 37P two unacc, textsyllables + i-ho-ha1J~/ = /22.22 321/,371 4 idem, 4371 /aihy aih/ = /22.32 1/'
Perhaps aihi is pUl?pa-form for ihi, the treatment of the second aihi in that
case is to be compared with AbC, typical is the treatment of the double
aihi of which thc first is unaccented, as is often the case in SV -ic melodization. The undulation in the vikrti-tones 2321 in 37P is also interesting.
Conclusion. The OCCUITence of the praIq-ti-tone 1 or of the combination 21 at the end of a saman-strophe is dependent on a text or sentence-stobha en ding in an udatta-syllable. Number of cases registered:
3 + (4 + 9) + (5 + 4) + (3 + 3) = 31. If we go out from the bhl?accentuation, we can formulate the rules:
bisyllabic parvan /31/ > /221/ (cases a and {l)
moresyllabic parvan / ... 1. 31/ > I . .. 1.121/ (case 15)
/ ... 2.31/ and / ... 231/>/ ... 2321/ (cases y, e, C)
Cases where a final parvan ending in udatta is melodized otherwise are
rare as will appear from the following subsections of the classification;
their explanation is difficult.
Il.
A. The final :r>arvan consists of sentence-syllables or a

sentence-sto bha.
The case is very rare. I have only registered 5541, textword vicak~ar,uiJp
with inserted stobha; /vic-uva/k~ii1}M = /33 31 , 1123/32342/. Probably the
first parvan is to be read /22 31 .11 23 /, owing to its very pecular treatment
the cadence may not be considered as an exception to subsection I Conclusion (supra).
61
B. The final parvan consists of or ends in a sound-stobha.

a. The sto bha is monosylla bic.
a. The stobha with a prakrti-tone 2 preceded by two unacc. textsyllabies: 176/ll- 23 2/, (here 1-2 used as a transcription of the vinata);
4381 /gr'fJ + / = /23.2/, 511 9 /-nyyalJ, + / = /23.2/, 547 1 /mdalJ, + / =
/23.2/. In 379 /-janat + / = /22.2 32 / the is sung on a figuration. The
stobha preceded by two syllables of a sentence-stobha, 469 8 fjvra +
+ / = /232/. No influence of accentuation apparent.
p. The stobha hi; one or two text-syllables precede; last vowel
(+ cons,) > 0: 1693 /235.2/, 3422 (ye-)/mire + / = /122352/, 446 (juja-)/
j~ate
b.
+ j = fI '22 45 2t/
The stobha is- bisyllabic.

Stobha irf,: 2375/123452/2/.
lIl.
Only two doubtful cases occur: 50 (a)-/dhvare + / = /233/, read /232/1,

258 1 , BI. I 533 sq. the so called saman consists only ofstobhas; last stobha
jsatYaSrvase/ = /212223/'
IV.
A. The final parvan consists of text-syllables or a sentence-sto bha.

a. The tone is a prakrti-tone.
The parvan consists of two text-syllables. P (bar-)/hi~i/ = /3 234 4/, 541
(kr-)/~tyalJ,f = /3 234 4/, 542 idem, 551 2 (mah-)/yuvalJ,f = /3 234 4/.
b. The tone is a vikrti-tone.
5804 (vana-)/prak~m udaprutam/ = /21.1213 234 / a> m, M; 511 14 sentence-stobha in the form of a trilil~ubh-pada /ti visvni duritd tarema/ =
= /12.122. 22l. 223 234 / a> m, no M.
B.
The two final parvans consist of a sound-stobha.
Stobha irf,: 61/134/4/, 162/1234/4/.

V.
A. The final parvan consists of textsyllables or a sentence-sto bha.

a. The tone is a prakrti-tone.
a. The final parvan contains two syllables of the text; its form is
/3 234 5/. As to the instanees I have limited myself to the first forty cases
and no notice is given to the textual accentuation except when it is in
contradiction with the rule given sub subsection I: 341, 56, 89, 91, 1002,
103, 104, ll71 , llga, 121 (dim), 1271, 1291 , 1321, 1325 , 1381, 1591, 1592, 16P,
173, 185, 213, 214, 223, 2365 , 245, 2481, 249, 251 2, 2521, 2751, 2752, 2792,
377, 3842, 3882, 3922, 3931 (divlJ,) , 4021, 403, 4051.
62
{J. The final parvan contains a bisyllabic sentence-stobha; its form is

/31.34.5/: 78 (visaM, 1593 (ih against I concl.), 2461 (vyaM, 2613 (abhi
against I concl.), 261 4 (di{JaM, 3821 (divaM, 3823 (kaM, 3892 (indral)" textword as stobha), 39Jl (dyJJhiM, 3923 (iti), 450 (dhnam), 4502 (dhrma),
452 (visaM, 4678 (gvbhil)" BI. II 523l. 7 gvmiM, 4698 (juva), 471 6 (diSaM,
482 2 (gvbhiM
y. Exceptional forms. Syllables precede tone 3: 370 stobha /i diva/ =
= /22.32345/, 51118 /hirar.tyyal),f = /3234.63234.5/. Between 3 and 5 only two
vikrti-tones: 5121 two unacc. textsyllables /334.5/; irregular series of vilqtitones 2483 (car{Jar.ti-) /dhrtil),f = /3 231 5/. Very outstanding 4392 stobha
slkal), /4234.5/. The so called sman 5251 and 5252 only contain the melodization of 525 p. a, so their last parvan (iraya-)/ti pr vhnil),f = /2345/
is not really the final parvan. For 525 vide below subsection VI.
(A)b. The tone is a vikrti-tone.
Subdivision : a-{J-form / .. . 12345 1; y-l5-form I . .. 3234./; irregular forms.
a. Form I . . . 12345 /, the final parvan contains as a rule two Or more
syllables of the text, rarely one. Specially marked are (Ily) the cases
where the last syllable of the parvan is udtta and wruch therefore go
against rule IConcl., and (2Iy) the cases which show an evident influence
of the accentuation on the melodization whilst fina! 12 is changed by
partial metalepsis into 112346' (cf. PS. 9,30 case 1, supra p. 41).
Instances: 681, 682, 74 dhanarcim, 771 tanpdl)" 1151, 1172 (hirar.tyyal = 12211 2345/, 1444, 1582 (arcan-)/tu lcarval),f = 12.2311234./, 163 (in-)dram tyel = 12 . 2311 234./, 1701 (a cyavaya-)/siy t-Iyel = 12.23111./.34.5/,
2381 (~te)/vasudruvaml = 12. 211234.5/, 3231 /adhad rdl),f, 3233 idem. 3271,
371 5 two last syllables repeated, 3834 1harisriyaml = 12211234./, 3881
/panasyvel = 12211234.5/, 3883 idem, 3981/ndrval = 11l234./, 4272, 4291,
4302 , 431, 4332, 435, 464/krpd sual),f = /21.11 / M?, 492 (de-)/vayu1!~
2346
jnam/ = 122 .1l234.5/, 4952 (na-)/vatir nva! = 121.11234./, 5112 , 51Jlo, 511 12
Ihirar.tyyal),f = 12211234./, sic COIT., 51215, 529/svan ldril),f = 121.1/1 234.5/,
542/ .. . jytir in-Idul),f ~ I ... 12.1/12346/, 545 (dirgha-ji-)/hviyaml= 111234.5/,
55Jllgre mahiyuval),f = /1l-12.2211~/, 563, 5662 two last syllables
repeated, 5663 two last syllables repeated, but melodized differently,
567 5 /suarvidaml = /2211234./, 5823/smolyl), lfUk..'litilndm/ with the last
o of somo changed into the stobha auviiova and with the syllable
ti changed into the stobha tauvaye = 11223234.5/112231123/12346/, 5832
(amrtatvd-)/ya gho{Jyanl = 12. 23 II 2346/ . .
The total number of .i nstances amounts to 42, of which 4 cases have the
udtta on the last syllable and are exceptions to the rule IConcl., and of
which 17 cases have the udtta on the penultimate syllable and COnBequently the dependent svarita on the last, cf. Simon p. 524, note on
PS. 9,30 fust case and supra p. 41. The remaining cases may be divided
as follows:
12 21234.5 in 681, 682
>
12)
22)
112)
122)
222)
222 )
63
2312345 in 511 2, 51120

2312345 in 51215 , 563.
2112345 in 4291.
22312345 in 1444, 427 2, 4302, 431, 433.
22312345 in 3271.
1112345 in 435.
(J-form: / ... l za.5/; the final parvan consists of a sentence-stobh.

Accentuation and melodization are both given.
Instances: 131 /8vii gdvaM = /123112345/' 491 /d~iiya/ = /12312345/' 711
/divam/ = /llz345/' 71 2 idem, 1282 /yyuM = /1l2345/' 141 /d~iiya/ =
= / 12312345/' 15()2 two text-syllables + stobha /sut1ft rayi~fMM = /21 .
. 121 2345 /, 1553 /lcaM = /2312345/' 1941 /yyuM = /1l2345/' 195/harisrtM =
= /22312345/' 2352 /subhtye/ = /223112345/ sic COIT., 2363 /bhgiiya/ =
= /223 12345/' 24P fjanitram/ = /2112345/' 2462 /vyobhiM = /22312345/' 2485
/sUar myaM = /11.112345/' 27P /su81ft8aM = /2112345/' 271 2 /su81ft8aM =
= /22312345/' 2723 /srvase/ = /22312345/' 320/diva?t, BI. II p. 523 1. 2/ =
= /1l2345/' 4292 /dhrma/ = /1l2m/' 4391 /slokyata/ = J21112345/' 4694 /asm
rdya ut srvaM = /21.11- 2z . 21.11 2345 /, 4703 two textsyllables + stobha
/sah rayil}tMM = /22.22312345/' 4762 /iym/ = /2312345/' 541 2 /didyat/ =
= /2112345/' 544 /vdhaM = /2312345/' 562 2 /divi/ = /lIz345/' 5623 idem, 582'
textsyllables as stobha /smaM = /ll234''/'
Ttal number of instances 29, of which 6 cases have the udtta on the
last syllable against rule I conclusion, and of which 14 cases have the
udtta on the penultimate syllable with the change of 12) 11 2345 . The
remaining cases may be divided as follows:
12 ) 2312345 in 1553, 544.
21 ) l1 z345 in 5622, 5623.
122> 12312345 in 491, 141.
) 22312345 in 2363, 2462, 2723.
312) 22312345 in 271 2.
3122 ) 21112345 in 4391.
Conclusion to ba and {J. Total number 72 ofwhich 10 with udtta
on last syllable and 31 with udtta pn penultimate syllable and regular
developIIlent. The 'remaining cases' have exclusively the tones 1 and 2
as prakrti-tones, cf. the introductory note to this chapter ad c (supra
p. 58).
Second case, (yand 15), the svrya begins with tone 3, and third case
(e), irregular cadences.
y. Form / ... 32345/. The fina.l parvan consists Of text-syllables. Instances: 12 /barhil}i/ = /22332345/' 42 (Bu-)/kr dhutaM = /21132345/' 53 /rul
vdyam/ = /rul vdiyiim/ = /21132345/' 62 mrtyasya (with pUlllpa and stobha)
) /mrtiiyii auhovJsyii/ = /llz2234555/3Zw./, II (ami)-/tram ardaya/ =
= /2.223 2345 /, 19 /vivsvabhiM = /2 II 32345/' 301 (rtnii-)/ni dii8~e/ = /2 .
. !H32345/' 52 unacc. syll.: (su-)/krato 'Pr".W/ = /22.232345/' 61 2 /ca vdriyam/ =
= /2.1132345/' 701 unacc. syll. aBoci (with pUlllpa and stobha) I ... siiu
64
2/23/32345/' 702 idem, 76 bhtv asm> /bhtu hau v/,smi/ =

= /4455/3 2345 /, 8P /vjya pnthm/ = /122.13 2345 /, 82 (bhak~ta dai-)
vyam/ /vym/ = /3 2345 /, 97 5 /mahsya/ = /213 2345 /, 98 /00 vedhse/ =
= /1213 2345 /, 1093 (havym uhi-)/~e/ = /3 2345 /, 116 (m)de madeJ;, with pu,?pa
and stobha /dim uv/del;j = /11-2323/32345/' 1242 (rari-) /m tel = /13 2345 /,
134 pda c /vsu sprh'f/t td bhara/ = /12.21.1.123 2345 /, sic COIT., 1443
(m'f/thi-)~tham/ = /3 2345 /, 1481 pda c /ttra ~ bhuvat sc/ = /12.21.
.22.13 2345 /, 1482 /bhuvat sc/ = /22.13 2345 /, 1671/d~i1Jena/ = /122332345/,
1672 idem, but without vikrti-tone 3, 171 /sani'f/t medhm aysi{lam/ = /21 .
. 1.21.1233 2345 /, 180 (abhi-)/{ltir jas/ = /2.113 2345 /, 1821 pda c, JindraB
crmeva rdasi/ = /11.11-222.1132345/' 189/dhiyvasul;j = /21132345/' 211
(-)/jaya sprdhal;j = /22.13 2345 /, 233 1 two last syllabls repeated (ta-)
/sthu~al;j /sthU{laJ;, sth~l;j = /22.13 2345 /, 2351 (sahsre1Je-)/va sik~ati/ =
= /2113 2345 / sic corr., 2361 unacc. syll.: (na-)/vmahe/ = /133 2345 /, 2362
navmahe> /nav/m/hi/ = /323/12345/32345/' 243 2 (dhr-)/~1Jum jas/ =
= /2.113 2345 /, 253 1 /crmasi/ = /11-2213 2345 /, 259 unacc. syll. /aBmahi/ =
= /2113 2345 /, 2881 (vi- )/vratnm) /vrt auhov/nm/ = /1 23 234 555/32345/'
3152 (yd dna-)/vn han/ /vii/y auhov/hn/ = /12/3234555/32345/' 322
/sthvirya tak~ul;j = /1122.23 2345 /.
Total number of instances 40. As a rule the last syllable bears a pracaya
accent which follows PS. 9, 30 third case (supra p. 41). In some cases a
v/c/
/ ...
>
>
>
>
>
sound-stobha iE, inserted within the reach of the last three syllables which
makes applying the rules uncertain: 62, 701, 702, 76, 116, 2881 , 3152. There
is only one case with udtta on the last syllable, 76, which, however, also
shows stobha-insertion. The number of cases with udtta on the penultimate syllable and thus going against PS. 9, 30 first case amounts to ten;
12,301 , 8P, 82, 97 5, 98, 124, 1481 , 1482,211, three ofthem may be explained
by the character of the last syllable, being a contraction of two syllabies:
8P, 82 and 97 5. Peculiar is the fact that the cadence obeys the general
rule, whilst at the same time one of the syllables on which the rule does
not bear is raised (2) 1): penultimate 2531, praepenultimate 2362, syllable
before praepenultimate 2361 , penultimate and praepenultimate together
259 - or is lowered (1 >2), syllable. before praepenultimate : 4 2, 180, 243 2.
15. Form / ... 32345 f. The final parvan consists of a sentence-stobha:
401 huve vsu, verb unacc. although beginning a sentence, BI. n 5191. 2,
/huve vsu/ = /22.13 2345 /, 402 /vid vsu/ = /22.13 2345 /, 641 four textsyllables + stobha /dty'f/t cran mah/ = /21.11.13 2345 /, 901 stobha of
eight syllabies: J.ynim indraB ca gacchathal;j = /11. 11. 2 . 223 2345 /, 1192
verb unacc. /d vsu/ = /2.13 2345 / sic corr., 1221 /agnir hutal;j = /21 .
. 123 2345 /, 1222 /sukr hutal;j = /21.123 2345 /, 1322 stobha of eight syllables
/asmbhya'f/tgtuvittamam/ = /212.22123 2345 /, 1382 /havi{lmate/ = /2113 2345 /,
141 2 = stobha 1322, 154/gva Mvl;j = /11 213 2345 /, 1653 /ghrtacutaM (BI.
n 520 1. 6) = /2213 2345/, 206/ti dvi~al;j = /12.13 2345/, 2484 /suvar mahl;j = /11.23 2345 /, 258 2 /srvase/ = /1132345/,2713 textsyll. repeated as
stobha I-sul = /3 2345 /, 2832 /stU{l/ BI. n 532 1. 1 = /2332345/' 321 2 /rtm
65
amrtaml = 121.1132345/, 3551madhuScutaM = 122132345/, 3723ldhrma1Jel =

= 11132345/, 3781 tri!l!~ubhpda: /{ndulJ, samudrm urviyd vibhtil = 111.
.221.121.123 2345 /, 3782 idem, 426 = stobha 206, 427 3 /svrvatel = 1112323451
4301 /vidharmal = 11232345/, 438 4 = stobha 4273 , 440 Isvratal = 122'332345/,
4653 stobha of eight syllables IviSvarp, sm atri7Jaf!/- dahal = 112.1.111.
.23 2345 /, 4654 similar to 4653 , 467 2 fvail = 13 2345 /, 47I3 = stobha 1322 ,
511lsadovi~aM = 1221323451, 5545 /vji jigivd viSv dhnnil = 121.121.
.12. 11 232345 /, 5651 lark devdnm pararn viyomanl = 121. 111 2 .221.
. 11232345/.
Total number of cases 34. In 18 ca.ses the last syllable bears a pracayaaccent and follows PS. 9, 30 third case, however in 11 cases the udtta
rests on the penultimate syllable against PS. 9, 30 first rule, and in 641
the last syllable bears the accent. Some cases have no significance as they
are monosyllabic : 27I3, 467 2 Sometimes there sooms to be uncertainty
about an enclitic verb: 283 2 , 440.
Conclusion to by and b. Although the rule of PS. 9,30 with
reference to the melodization of a pracaya-syllable prevails so that the
svrya bgins with 3, yet the number of cases in which one would expect
the rule for the svarita-accent to be followed, are rather numerous.
e. Irregular forms. - 232 Iv sta~el = 13322 345 1, 46713 Ivlsta~el =
= 123/222345/, 531 stobha + one text-syllable Ihiyhuv sthatl = 1556.55 .
. 52345 /.
(V)B.
The par van consists of or ends in a sound-stobha.
a. The sto bha is monosyllabic.

Instanees are numerous, but often of little interest theoretically; only
a few examples of each kind will be given.
Sto bha , I ... 12345 1: 1572 , 547 4
Stobha i: when the parvan contains only i, its form is 132345/,672,721,
722 , 97 1 , 97 2 , 1154 etc. When the i is preceded by two text-syllables, the
parvan runs 1223123451: 4603 , 2522 , 371 6 , 547 5
Sto bha . Rarely occurring, form 13 2345 1: 1091, 2453 , 3822 etc.; preceded
by two text-syllables 2393/2232345/.
Sto bha e preceded by two text-syllables 1223123451: 182 , 1681 etc. or
12232345/: 3824, 5726 ; two stobha-syllables' + e in 467 2 12232345/.
Stobha v often met with; six forms:
a. I ... au ho v/v/ = / .. . 555/3 2345 1 in 458.
p. /v hu v/ = /223112345/ in 270.
/three text-syll. + hu v/ = 14455.55/ in 75.
/throo syll. of sound-stobha + hu vl = /5556 55/ in 2452
y. preceded by hu in penultimate parvan:
/hOi/hi/auho auha v hulv/ = /1/1/21.21.123455/5/ in 20.
/uhuv hu/vl = 15556 .5/5/ in 1333 .
/hu hova hu/v/ = 12.223.2/12345/ in 534.
b. preceded by u in the same parvan ; four text-syllables of which
66
the final syll. has vowel (+ cons.) >; /211231.12345/ in 5551;

variant: three text-syllables: /222322345/ in 5553.
B.
preceded by u in penultimate parvan ; there are three patterns,
I give the text-syllables here in italics :
(magh-)v/n amtye/ /nmtay au/vii/ = /211231/12345/ in 3752.
/ccha/ >/acchu/ vii/ = /452/22345/ in 543.
/priy1ft mitr1ft n 8arJUli~am/ /prya1ft mitra1ft na 8amsi~au/vii/ =
= /11.11.1.1 232/2345 / in 354.
1;. These patterns actually form a special case of B. A number of syllables (often four) which close the verse-strophe, is taken to form a
parvan; of the last syllable the vowel (+ cons.) is changed into u.
rhe first syllable or syllables are sung on tone .4, then follows one sung
on 45, then follow one or two syllables sung on tone 5. The stobha v
forms a separate parvan on tone 5. To this group belong: 4661, 4686,
4971, 5494, 5552; 3801, 4571, 536, 5371, 5372. So e.g. 4661:
/viddi~am/ >/vididi~au /vii/ = /44. 555/5/.
Stobha he, /3 2345 /: 267, 5145.
St ob ha ooi, very often occurring; the last syllable of the preceding
text has its vowel (+ cons.) changed into o. Three main patterns:
Ist pattern: /123456.5/ : IS, 51, 71, 72, 1702 etc.
2d pattern: /234556.5/ : 52, 982, 992, 105 etc.
3d pattern: /4 556 .5/:43, 81 , 9, 2Petc.
>
>
(V B) b The stobha is bisyllabic.
Sto bha irfii, /45/ : 184. Cf. oirfii and hOirfii.

Sto bha iti, /3 234 5/ : 21 2, 1181 , 172.
Sto bha iha, three patterns:
lEt pattern: /3 234 5/ : 1203, 1302, 1332 etc.
2d pattern: /2345/ : 218.
3d pattern: /13 2345 / : 1552, 340.
or: /23 2345 / : 34IS, cf. 37F.
Sto bha ihi, 2463 : /3 234 5/ and 505: /232345/.
Sto bha upa. Often occurring stobha. Patterns with tone-movement
on the first syllable: u. /3 234 5/ : 331 , 332, 92 etc., (J. /2345/ : 1431 , 231, 2551
etc. y. /223234 5/: 1551, 1601 , 227 etc. - Patterns with tone-movement on
the second syllable. (J. /132345/ : 230, 318, 4754 etc., B. /12312345/: 1451,
1452. 1;. /12 2345 / : 363. - Pattern with no movement on either syllable:
4023 : /3234.5/45/ with upa repeated.
Sto bha oirfii, very often occurring with a great number of special
forms. The pattern /oifrfii/ = /12345/5/ and /textsyll. + oirfii/ = / ... l2345/5/.
Instances of /12345/5/: 3, 41, 82, 14 etc. etc. Instances of /11.12345/5/: 4101;
last textsyll. with vowel 0: 5133: /123212345/5/. - Pattern /223234.5/5/: 2821,
2822, 3843 etc. final vowel (+ cons.) 0: /33 2345 /5/ : 35P, 3924. etc. Pattern / ... 2345/5/: 152, 4802. Pattern / ... 42345/5/: 44P. Pattern
/145/5/ : 32.
>
67
Sto bha hirfii" very often occurring. General type /45/5/; special types
/12345/ 5/ and / ... 12345/5/, /3 2345 /5/ and / ... 32345 /5/.
Type /45/5/ : 2, 12, 16, 171, 173 etc. etc.
Type /12346/5/ : 66; type / ... 12345/5/ : 456 2
Type /3 2345 /5/ : 343 6, 343 7 etc. etc., type / ... 32345/5/ : 575.
VI.
The last syllable always has the figuration 5668 . Only syllables of the
text are used. According to the tones we can distinguish three types
/ 123455656/' /23 455656/ and / 32 345 5668 /.
/123455656/ : 396, 5461 , 5546, 560.
/23455656/ : 65, 69 (sic corr.), 80 etc. etc.
/323455656/ : 23 1 , 232 , 342, 123, 162 etc. etc.
GENERAL CONCLUSIONS TO THE SIX SUBSECTIONS
Looking over the material given in the classification we can state that
there are the following general cases:
A. The chant-strophe ends in tone 1. This ending is strictly dependent
on the text, its condition being that the last syllable of the text bears
the ' udtta.
B. The strophe ends in tone 2 or 5. Tone 2 is the harmonic substitute
for 5. Tone 2 specially occurs in stobhas: ., hi, irfii,. Tone 5 is the most
common of final tones.
c. The strophe ends in 3 or 4. In the grmegeyagna only doubtful
instances of 3 occur. We could speak in the case of those cadences of an
'interrupted' cadence, it is as if the aim is not reached.
D. The strophe ends in the final tone 6. The cadence, as it were,
overreaches its aim.
Next to the question of the pitch of the final tone we have to do with
its function as prakrti- or vikrti-tone. Since the last prakrti-tone of a
parvan is specially stressed, we may call the use of the final tone as prakrtitone the masculine ending and its use as a vikrti-tone the female ending.
Tone 6 is only used in a female ending. Tone 1 is used in both functions
and in both cases the parvan to which it belongs, contains exclusively
the tones 1, 2 and 3. Tone 2 as final tone mainly occurs in stobhas and
with one exception (379) it is always prakrti-tone; further we notice that
it is either preceded by the adjacent tone 3 or reached with a leap from
tone 5. Tone 4 is either prakrti-tone and reached through the cadence
/3 23 4/, or vikrti-tone with the two cadences /3 234 / or /1234/. Tone 5, represented by a great many instances, is either prakrti-tone: /32345/,orvikrtitone in the two normal cadences /3 2345 / or /12346/' sometimes 5 as prakrtitone forms a separate parvan in sound-stobhas, vide sub V B v, irfii, and
hirfii" in this case it is the repetition of a preceding prakrti- or vikrtitone 5.
68

SECTION II
ANALYSIS OF THE SAMANS ENDING IN 5 U8
The samans ending in the cadence 5858 form a small group containing
three subgroups. In describing these samans I shall mention the place
which they have in the :.;tgveda-SaIphita, the gods to which they are
devoted, the metre in which the textis written and, if desirabie, the
irregularities found in the metre. Then I shall state whether the syllablecontractions are dissolved in the chant, and the numerical relations in
which the text-syllables are distributed over the parvans.
First subgroup: /123455858/. It is only represented by four instances.
Gramegeyagana 396, :.;tV. 8, 24, 24. Indra.
Pragatha 8 + 8 + 12; p.da a contains seven syllables instead of eight,
the genitive ending -m pot the end having the value of two syllabies, the
melodization has left the number of seven unchanged. ~o stobhas are
inserted. The melodization is based on the accentuation. Syllable-division:
a: 8 (7), b: 7 + 1, c: (3 + 1) + 4 + 4. For comparison both the numerals
of accentuation and melodization are given. The digit 2 as substitute of
1 by special Samavedic law of sound is printed in italics.
Melodization
/45.4.4554/
Accentuation
/23.1.1223
122~. 2312/
1222.31. 2312. 22/1
/1122.221/1/
/122/1./21.22/123.455658/
In pada a, evidently forming the prastava, the tones corresponding

to the accents are replaced by their fundamental tones; the udattas of the
first and fourth syllables are not extended over the following syllable, the
last syllable of pada a on the other hand is unexpectedly raised, cf. the
last syllable of pada band the fourth of c. In p.da b we find the acoustic
metalepsis of the two udattas, the eighth tone has become a separate
parvan. Pada c by way of climax has been divided into many parvans.
Gramegeyagana 5461, J;tV. 9, 101, 7 Soma.
4 x 8 syll. Pa. d has seven syll. instead of eight in consequence of via
contracted into vya; in the melodization the contraction is dissolved.
Syllable-division: all padas divided into 4 + 4 syll. In pa. a and c the fifth
syll. has its final V'Owel replaced by the stobha auho so that the parvans
aa and ca contain five syll. The melodization is independent ofthe accentuation. Melodization:
/22111/2111/
/22223/ 123455858/
/22111/2111/
/22223/ 12345 5858/
Melodically pa. a
pa. c, and pa. b
pa. d. Tht;) parvans aa and ap,
69
Ca and cfJ have the rising form, ba and da the monotonous form, bfJ and dfJ
the typical descent from tone 1 to tone 6.
Grimegeyagina 5548, :J.l.V. 9,75, 1 Soma.
4 x 12 syll. Pi. c with 11 syll. owing to the contraction rya in 3d syll.,
undissolved in the melodization. The fust two syll. of pi. a have become
three by a stobha: abhi abhyov. Syll.-division, a: (3 + 6) + 4, b: 8 + 4,
c: 7 + 4, d: 4 + 4 + 4. Melodization:
>
/544/211 111/221121
/1111 1111/22112/
/1111 111/221113/
/12345/221123/1234,,5868/
In parvan aa the tones (211) are replaced by their fundamental tones,
parvans aa and afJ monotonous with their beginning marked by tone 2;
ba and ca entirely monotonous ; ay, bfJ, cfJ and dfJ with rising movement
which ende in falling vilqti-tones; pi. da in its descending movement
preparatory of thc typical cadence in pi. dy.
Grimegeyagina 560. :J.l.V. 9, 70, 1. Soma.
4 x 12 syll. Pi. b with contraction vyo in 9th syll. dissolved in melodization; pi. c with contraction rya undissolved. In pi. a at the beginning
of the third parvan, before the llth syll., the stobha ho is inserted.
Accentuation of pi. a without stobha: 1.22.32.312.2222. Melodization:
/113&" /45. 54.545. 55/445/
/21 111 111/221121
/2 111 111/221113/
/223134.5/2112/1234,,5858/
In the parvans afJ and y the tones are replaced by their fundamental
tones, undulating movement, perhaps a m with M. In ba and ca monotony with marked beginning; in bfJ and cfJ rising four-syllabic movement
with descent in vilqti-tones at the end, da in its falling movement from
2 to 5 preparatory of dy with its typical cadence.
The last three melodies are cognate, the parvans use only two adjacent
tones except the last parvan of the strophe and the parvans which prepare
the final cadence: bfJ in 5461 , da in 5548, da in 560.
All the four instances are characterized by pdas containing an even
number, 8 or 12 syllabIes.
>
Second subgroup: /234,,5868/. lts classification is very intricate and

will be given in the form of a tabie.
70
A
1
a
a
{3
y
{3
3
The penultimate parvan has the form /2343/'

The figure /2345/ is met with at the end of one or more
pdas.
The strophe consists of 4 x 11 syllabIes.
The figure /2345/ is met with at the end of the first
three pdaEl,
at the end ofthe second and third pdas,
at the end of the second pda.
The strophe consists of 3 x 8 syllabIes. (Instanees of
this group are rare).
The figure /2345/ is met with at the end of the first pda.
The figure /2-11232/ is met with at the end of one or more
pdas.
The figure /2-11232/ is met with at the end of the first
three pdas.
at the end of the second or third pda.
The figure /2345/ is met with at the end of the second, and
the figure /2-11232/ at the end of the third pda.
The penultimate parvan has not the form /2343/' The two
smans belonging to this type are very differently constructed, but prosodically they both have pdas with an
even number of syllables (8 + 8 + 12) and 4 x 8. Cf.
article Ab.
A 1 a a. - Penultimate parvan : /2343/; strophe 4 x 11 syll.; figure

/2345/ at the end of the first three pdas.
GG. 65, :.;tV. 10,56,1. Agni (ofcremation). SyIl.-division, a---c: (4+3+4),
d: 4 + 3 + (1 + 3). Stobha-line before pda a and d. Melodization:
/oh/hahi/ = /3 45/32/
/2111/2311/2345/
/2111/1123/2345/
/2111/211/2345/
/oh/hahi/ = /3 45/32/
/2111/2311/2343/23455858/
GG. 69, :.;tV. 4, 3, 1. Agni. In pdas b-d the four beginning syll. are
divided over 2 or 3 parvans. In p. a and c within thegroup formed by
tlte 5th-7th syll. tone 3 is used as a prakrti-tone. No stobha.
GG. 80, J;tV. 10, 87, 19. Agni as a killer of Rak~as. Melody much resembling that of 65, but other stobha-lines.
GG. 3252, J;tV. 10, 55, 5. GG. 330, J;tV. 7, 23, 1; both to Indra.
GG. 339, J;tV. 10, 89, 4. Indra. Pda a repeated in unchanged form.
Stobha-lines between th~ pdas.
71
GG. 523', ~V. 9, 78, 1; GG. 524', ~V. 9, 97, 7, both to Soma.
GG. 525, ~V. 9, 97, 34. Soma. The so called smans 5251 and 5252
contain only p. a five times repeated and preceded by different soundstobhas and 5253 contains the pdas b--4, again with different soundstobhas.
GG. 526, ~V. 9, 97, 1. Soma. 526 has the syllable-division: a-c
(2 + 2) + 3 + 4 syll., pda a is twice repeated; p. dis first treated similarly to p. a-c and then its last six syll. are repeated and grouped as
2 + 1 + 3. The first parvans of the pdas are increased with the trisyllabic
sound au/wvii. Melodization:
a
13234 , 3451 11/2311j22345/uhuvi = 3321
14234.3451-1 -
13234 .345/-1 -
I -
Ihi
2/uhuv = 222,1
12345/hu = 2/hi = 21
I - lh = 5/h auhi = 2221
=
eli-I d
I
1-123311 - I
1-/ 2343/ 234558MI
The second form of 526 is simpier in its construction, it has only stobhalines before pdas a and d, does not r.epeat pda a nor a part of pda d
and neither inserts stobhas within the pdas.
GG. 527, ~V. 9, 96, 5. Soma. lts first chant-form will be discussed
sub A 2. The 3d and 4th chantform are nearly equal in the melodization,
they differ specially in the stobha-lines before the pdas a and d.
GG. 528. ~V. 9, 90, 2. Soma. In p. c the daJ).la has been omitted after
the second parvan. - GG. 532, 1.tV. 9, 96, 13; GG. 533, ~V. 9, 96, 1; GG.
535, ~V. 9, 97, 4; all to Soma.
A small group of smans belonging to this group is characterized by
the syllable-division of the pdas a-c : (2 + 2) + 3 + 4 and of d: (2 + 2) +
+ 3 + (1 + 3).
GG. 3131 , ~V. 7, 21, 1. Indra. The first parvan of two textsyll. adds the
stobha au/wv Mi at the end, before which the last vowel (+ cons.)
is dropped. Melodization of the first two parvans 1222232/323,5/.
GG. 3141 , 1.tV. 7, 24, 1. Indra. Parvans a and {J in all the pdas: 145/11/.
GG. 3162, 1.tV. 10, 148, 1. Indra. The same sOlmd-stobhas precede the
four pdas. Parvans a and {J in all the pdas: 132345/11/.
GG. 5231 , ~V. 9, 87, 1. Soma. Sound-stobhas inserted between the
parvans a and {J. A monosyllabic stobha is put before the two syll. of the
first parvan, e.g. pda a: /i-pratu = 235/ih = 3231 /drava = 21/.
One sman is still more em bellished :
GG. 3251 , ~V. 10, 55, 5. Indra. In the second parvans of all the pdas
the two syllables are repeated. Melodization of the parvans a and {J:
145 / 11 2- 11 /.
A 1 a {J and y.
One and the same sman fills the two articles of the classification. I
72
shall fust give the text with its lldttas marked and then the two forms
of melodization:
00. 332, ~V. 10, 178, 1. Trk~ya.

tym u su vO,jina/,,,, devjuta1!'
sahovdna1!' tarutdra1!' rthiiniim/
ri~tanemi1!' prtandjam 8U1!'
svastye tdrkMfam ikd huvema/ /
Melodization of 3321 with the figure /2345/ at the end of the 2d and
3d pdas; the figure 2343 , which regularly forms a separate parvan, is
here placed at the end of the penultimate parvan:
22.223.2/2.2.2.2223.43335/ .(the penultimate syllable should be accented 4)
/iya iy = 345. 45/hi = 3834/5532.1/223/2.345/
/iya iy hi = 22.223.2/2223/1/23.11/23.45/
/iya iy = 345. 45/hi = 31345/21/1./12.32343/23455858/
/iya iy hi
Annotation. The preceding stobhas ofthe pdas a and care characterized

by monotony on tone 2, those of the pdas band d by falling movement.
Pda a has no syllable-division; b is irregularly divided into 5 + 2 + 4
syllabIes, pda c normally into (3 + 1) + 3 + 4, and pda d again irregularly into (2 + 1) + 4 + 3; the contractions at the beginning of tyam
and svastaye are not dissolved.
Melodization of 3321 with the figure /2345/ at the end of the second
pda; the figure 2343 forms again the end of the penultimate parvan in d.
/5.5.5. /22/2345' /2223 134 /
/5532.1/223/ 2 . 345 /
/2223 234 5. /222343 2 . 35/
/21/1. /12.32343' /23455858/
All the final parvans of the pdas are characterized by falling movement.
A 1 b. The verse consists of 3 x 8 syllabIes. The figure /2345/ is

found at the end of the first pda.
00. 196, ~V. 10, 105, 4. Indra. Metre irregular. In the melodization
the three last syllables of pda band the three first of c are taken together,
whilst there is moreover a superfluous syllable in this parvan:
/45. /2223' /234. 5/
/22.223/112.
1.3.21 83 , /32343/2343/23455858/
00. 2282, ~V. 10, 105, 1. Indra. Metre very irregular; syll.-division of a:
2 + 2 + 2 + 4, of b: 2 + 2 + 4; the first parvans of all the pdas are
increased with the stobha auhovO,:
/3234.345/11/223/2345/
73
/12 34 .345/11 2/32 232345/

/3234.345/1113/32343/23465858/
A 2 a. The strophe consists of 4 x 11 syll. The

with at the end of the fust three padas.
GG. 333, J;tv. 6, 47, 11. Indra. Melodization:
fig~re
/.-11132/ is met
/2111 I1I1/z-11132/
/1111 1111/2-11132/
/2111 111I/z-11132/
/212/1/211/2343/ 23465858/
A 2 {J. Strophe of 4 x 11 syll., figure /2-11132/ at the end of the

second or third pada.
GG. 319, J;tV. 10, 73, 11. Indra. Syll.-division a = 2 + 9, band c undivided, d = 3 + 1 + 3 + 1 + 1. Parvan a increased with the stobha
hiihiiu; the stobha-parvan /auhi/ = /,,-1323/ is inserted between the fust
two syllable-parvans of d.
/55.55/1111 11112/
/2111 1111 z-11132/
/1111 1111 21132/
/221/ z-1323/321/2311/23'3/2345658/
GG. 5271, J;tV. 9, 96, 5. Soma. Melodization:

/2222 2222 23112/ .
/2222 2222 23112/
/22 111 I1I/z-11za2/
/2211/2311/2343/23465858/
A 3. Strophe of 4 x 11 syll., figure /2345/ at the end of the 2d and

/2-11232/ at the end of the 3d pada. Syll.-diviBion a = 2 + 3 + 3 + 3,
b = 4 + 3 + 4, c = 8 + 3, d = 4 + 3 + (1 + 3). Parvan aa increased
with the final stobha hohai.
GG. 67 1 , J;tV. 6, 7, 1. Agni. Melodization:
/54.45/231345/111/1232/
/2111/211/2345/
/21 111 111/2-11z32/
/2111/2311/2343/23465658/
B. The penultimate parvan has not the form /2343/' The two samans
belonging to thiB type are very differently constrncted; prosodically,
however, they both have padas with an even number of syllabIes.
GG. 974, :J;tV. I, 150, 1. Agni. Syll.-division: a undivided, b = 3 + 1 + 4,
c = 2 + 2 + 4 + 4. Melodization:
/552345445/
/211/1/1 t;.3 za45/
/31345/31345/22313,5/234:;5658/ .
74
Pda a belongs to the undulating type, parvans ba and {J are together

monotonous with beginning marked; all further pdas belong to the
descending type.
GG. 5493, 1;tV. 9, 98, 1. Soma. All the pdas are equally divided, and
only d differs in its melodization from the others. After the first two
syllables of all the pdas the stobha Mi is inserted in the same parvan,
and before the seventh syllable the stobha h.
Pdas a-c: /2221/222343/2232345/
Pda d:
/2221/222343/23~55856/
Third subgroup: /323455656/' All these smans are characterized by
a versification in which the pdas have an even number of syllabies:
3 x 8" 4 x 8, 4 x 12, 8 + 8 + 12 and 8 + 12 + 8. The classification is here
given in the form of a tabie; in a three-pdic strophe the first two pdas,
in a four-pdic strophe the first three pdas will be called the secundary
pdas. The smans are mentioned immediately under the article of the
classification, with their number of syll. between brackets.
A
I
Smans with the secundary pdas ending in tone 5.

The last parvans of the secundary pdas end in the figure
/ ... 3234 5/:
231 (3 x 8), 556 (4 x 12), 3851 (8
+8+
12).
The last parvans of the secundary pdas have the form /43235/:
342 (3 x 8), 557 1, 557 5, 56P, 5642 (all 4 x 12); 5791
The cadences / ... 3234 5/ and / ... 43235/ both occur in the same
chant-strophe:
362 (4 x 8).
The cadence / ... 3234 5/ is found next to / ... 555/ or / ... 5585/:
507 (3 x 8), 553 (4 x 8), 353 (metre uncertain).
All the secundary pdas end in tone 5, either prakrti- or vikrtitone, but none has the form / ... 3 234 5/ or /43235/:
5572 (4 x 12).
The last parvans of the secundary pdas end in one of the
higher tones (1, 2, 3):
5558 (4 x 12).
Note. Of 25P (8 + 8 + 12 + 8) pda a ends in the vikrt;i-tone 4,
pdas band c in the vikrti-tone 3. Sman 5555 (4 x 12) has an
intricate construction.
One or more pdas end in tone 5, and other pdas in one of the
higher tones.
The pda en ding in 5, ends in the figure / ... 3234 5/:
5574 (4 x 12), 5725 (8 + 8 + 12). For 3422 and 4632 see
the description infra.
The pda ending in 5, ends in the figure / ... 43235/:
(8
C
I
12
4002 (8
+ 8).
+ 12 + 8).
75
The pda ending in 5, ends in / ... 555/, / ... 5585/ or / ... 5545/.
1622 (3 x 8); 123 (3 x 8), 366 (4 x 8); 467 8 (3 x 8),

561 2 (4 x 12).
Note. On sman 123 (3 x 8) see the description.
The last parvan of the secundary pdas ends in / ... 32345/.

Instanees: 231, ~V. 4, 9, 1. Agni. 3 x 8 syll. Melodization:
A 1.
/112112/32232345/
/1111 2/323 2345/
/111123/ 3234 65868/
556, :!;tV. 9, 77, 1. Soma. -
3851,
~V.
8, 24, 16. Indra.
A 2. The last parvan of the secundary pdas has the form / ... 43235/.
Instanees: 342, ~V. 8, 84 (73), 7. Agni. 3 x8 syll. The same soundstobha precedes all the pdas: /auhihuv = 2223132/huva e = 232/,
pdas a and b: /2222343235/

pda c:
/2222/323465656/
5571, 557 5, ~V. 9, 86, 16; 56P, :!;tV. 9, 85, I; 5642,
:!;tV. 9, 108, 9. All to Soma.
~V.
9, 86, 43; 5791,
A 3. The cadences / ... 323,5/ and / ... 43235/ in the same chantstrophe.
Instanee: 362, ~V. 8, 69 (58), 8. Indra. 4 x 8 syll. Melodization:
/2222 223 23,5/
/22223 43235/
/2111 23 43235/
/111123/ 3234 55856/
A 4. Cadence / ... 323,5/ next to / ... 555/ or / ... 5565/.

Instanees: 507, ~V. 9, 47, 1. Soma. 3 x 8 syll. Melodization:
/5433244 555/
/2111/ 2-11 2323,5/
/122123/323465656/
553,
~V.
9, 101, 13. Soma. 353, cf. BIoomfield, Concordance.
A 5. All the secundary pdas end in tone 5, as prakrti- or vilqti-tone,

but the two ordinary cadences are not met with.
Instanee: 557 2, ~V. 9, 86, 16. Soma. 4 x 12 syll. In the fust three pdas
the two last syllables are repeated. Influence of the accentuation on the
melodization is traceable.
/344. 34. 345. 5/32. f23 2345 /
/32 3123,. /45.4.4555.5/4545/
/32 3123,. /45.5545.45/4545/
/32 3123,. /445.545/323.4555856/
76
B. The final tone of the secundary pdas is one of the higher tones.
Instance: 5558, ~V. 10, 79, 1. Soma. 4 x 12 syll. The two syllablecontractions in p. a are d8solved in the melodization.
/55/v hi
45/11.1.111/221 2 2/
/1111.1111/2.211~/
/1~2. /123/234/3.434.45/232/
/1 23 , 2/1~. 234/4.345/3.23.455858/
CLOne secundary pda ends in / .. . 323451; the others in one of the

higher tones.
Instances: 557', ij,V. 9, 86, 16. Soma. 4 x 12syll. Pdas a and c have
the same melody, and band d are almost equal. Melodization:
/2211/2211/ 2111 23/
/32 235/32235/323 234 5/
/2211/2211/211123/
/3235/32235/3234558581
5725,
9, 106, 10. Soma. 8 + 8 + 12 syll. The first pda ends in

the second in 12112/, the 7th and 8th syllables of pda c run
~V.
I ... 3~,5/,
13~,5/.
3422, ~V. 1, 10, 1. Indra. 4 x 8 syll. The pdas b, c and d are repeated,
vowel (and consonant) are elided before the stobha ohiii.
a:
b:
c:
d:
first time
15550hai = 45/32 23 234 51
11111 21232/
12111 121232/
Illll 1212321
second time
/222 oha = 32345/32232345/
/222 ohi = - /3223234 5/
/222 ohi = - 1323455858/
4632, ij,V. 9, 111, 1. Soma. In the rcika-reading the verse-strophe consists of seven pdas with the following number of syllabies, a "= 12, b = 12,
c-e = 8, 1= 12, g = 8. The chant-strophe 4632 consists of the pdas
a-c; the contractions in a and c are not dissolved. The pau~pika diphthongs i are here written by Satyavrata Smarami as yi, a less desirabie
spelling. In pda b the verb is read as tarasi and the final i is elided before
the stobha auho/v. Melodization:
/55.32.345/2211
/1l.1232/111/sauho = 22/v hi = 2323/112323451
12123.2/3234556581
C 2. This article of the classification (use of the figure I ... 43235/)

contains only one sman: 4002, ~V. 8, 21, 9. Indra. The original strophe
has the form 8 + 12 + 8 syll. Pda a is repeated; of pda b the last eight
syllables are repeated; in pda c the n of indram is dropped according to
pu~pa-rules :
a.
a.
/5.4 3.2344.551
/1.1.1112 1.2321
b.
b.
c.
77
/111.111/2.23.4 3 .2.35/
/11/2.23.43 .2.35/
/2112' /IS3/3. 234,,5856/
3. The pda ending in tone 5, ends in / ... 555/, / ... 5585/ or / ... 545/.
1622, :!;tV. 8, 82 (71), 7. Indra. 3 x 8 syU. Pda b is repeated. Melodization:
a.
/5555 58555/
/1111 121232/
/22223/231232/i = 1'2/
C.
/2223 234 hi = 5/3234:;5856/
123, :!;tV. 8, 2, 25. Indra. 3 x 8 syU. Pda a is repeated. Two cadenc61i1
of the secundary pdas occur, viz. /5585/ and /2345/. Melodization:
b.
b.
a.
a.
[45554[55 8 5/
b.
/1[223[2345/
C.
/112/123/3234,,5856/
/1111 121232/
366, cf. ~V. 5, 38, 1. Indra. 4 x 8 syU.; of the pdas b-d the four last
syU. are repeated. - 467 8 , :!;tV. 9, 61, 10. Soma. 3 x 8 syU. - 561 2, :!;tV.
9, 85, 1. Soma. 4 x 8 syU.
Conclusion to section 11. As the most important fact of this
section we may consider, I think, the way in which the three forms of the
5856 -cadence are connected with the versification and consequently with
the distribution of the syUables over the parvans. The form /1234,,5856/
and /3234,,5856/ are used in strophes with pdas of an even number of
syUables. The form /234,,5858/ is found in trililtubhs. The two forms/1234,,5856/
and /3234,,5856/ may he compared with the two cadences /1 234,,/ and /32345 /,
cf. supra pp. 62 - 65.
Further it is worth emphasizing the fact that the Smavedic chanters
have exploited their simple and limited outfit of means to the widest
range of possibility. In this respect their chant is cognate in spirit with
the Vedic ritual and poetry, one might caU it an art of the kaleidoscope;
however, one should not forget that this Vedic culture has been the
introduction to a later widely and deeply developed civilization.
SECTION III
OLDENBERG'S RULE OF ACOUSTIC METALEPSIS
Oldenherg's rule of acoustic metalepsis, i.e. the extension of the udtta

over the foUowing syUable provided that the second foUowing syUable
does not bear an udtta, is brought forward by the author with a certain
reserve (Prolegomena p. 485 sqq.). It seems to me a beautiful discovery
and its validity is completely proved by PJ.1n's rules for the Subrah-
78
ma~y (A~~dhyya
1, 2, 37 sq.). In my classification of the smans

contained in the Grmegeyagna this principle has also been of great use
to me. However, we have to accept that it is not a hard and fast rule.
It seems inconsistent to me that Oldenberg should quote hls examples
from the hyagna and the Grmegeyagna; it would be more correct
to use the Ara~yegeyagna together with the Grmegeyagna. And
besides it would, in my opinion, be more convincrng if the smans had
been written down in their melodie form. What I, then, wish to do is:
Ily, to write out Oldenberg's examples fully with references to the Grmegeya- and Ara~yegeyagnas, and 21y, to add to them a few examples
which I have come across in my reading.
The following abbreviations are used: GG. = Grmegeyagna; AG. =
Ara~yegeyagna; hG. = hagna; hyG. = hyagna; pA. = Piirvrcika; UA. = Uttarrcika; As. = Ara~yaka-Sarp.hit.
Oldenberg's examples from the hyagna.
Oldenberg p. 486. BI. V 387, hyG. I, 1, 7, AG. 1, 1, 3, text uA.
212 = ~V. 8, 70 (59), 5. Accentuation and melodization have been discussed in the Introduction, supra p. 21.
Oldenberg p. 486. BI. V 402, hyG. I, I, 20, AG. 3, I, 21 (BI. 11 454),
text uA. 460 = ~V. 10, 157, 1.
Accentuation
32.2.3.122.2223
12.3.12.2.32/
Melbdization (hyG.)
/hih = ll/21. 1. 1. ll2.22223/
/ih = ll/I1. 2 .11. 2.223/ etc.
The AG.-melodization agrees with this, but has more extensive stobhas.
From the hyagana we can leam that all the strophes are dependent on
the accentuation of their texts, and moreover that the final syllable is
independent of the accentuation, being everywhere melodized as 23'
Oldenberg p. 487 BI. V 408, hyG. 1, 2, 6; the sman is sung on the
yasas-melody AG. 2, 2, 12. The text of hyG. I, 2, 6 is uA. 663 = ~V.
9, 107, I, and the text of AG. 2, 2, 12 (BI. 11 441) is pA. 270 = ~V. 7,
32, 16. All the stobhas of AG. 2, 2, 12 have been omitted in the hyG.
melody with exception of the final sentence-stobha: satm jivema sardo
vay1Jl te.
Text of pA. 270,
Melody of AG. 2, 2, 12
(except additional stobhas)
(8 + 8 + 12 + 8 syll.)
11.2232.23
/ll.I22I.12/
1.222.232/
/1.122.221/
31.122.2312.223
/21.ll2.2311. 222
1.2.3.12.222//
/1.1.2.11.222/
stobha: /21. 122.212.21. 323t5/
>
In the verse Oldenberg's rule is applied : 122 ll2. In the stobha,

however, sardo-va- is melodized as 212.2.
In the following exposition of the accentuation of uA. 663 I have
79
dissolved the contraction of apsu antr ti in order to avoid the kampa,

which is permissable since there neither is a contraction in the chant.
Accentuation of U. 663
231.222.32
23.1.232.32/
331.1.12.32 . 32.2
323.23.122//
Melodization of hyG. 1, 2, 6
/121.122.21/
/12.1.121.21/
/221.1 . 11.21.21.1/
/212.12.112/
stobha: /21.122.212.21. 32345/
To follow the yasas-mewdy here means to base the melodization on the

accentuation with appliance of the accoustic metalepsis under the prescribed condition, and further to use the samefinal sentence-stobha. The
foregoing is astrong proof for the thesis that for the sman-chanters to
follow the spoken accentuation meant a type of melodization.
Oldenberg p. 487. BI. V, 433, hyG. 2, 1, 12, AG. 6, 1, 15. Text S.
5, 2 = U. 803 = :!;tV. 10, 170, 1. 4 x 12 syll. The stobhas are omitted
in the following:
Accentuation
32 . 31. 222. 32.23.
13.12 . 3123.1222/
1223.1.23123.12.
31.222.231.1222//
Melodization of AG. 6, 1, 15
/21. 21.122.21.11/ (sic corr.)
/12.11 . 2112.1122/
/1122 . 1. 12112.11/
/21.122.221.1122/
hyG. 2, 1, 12 has the same melody(mahdivkrtyam) as AG. 6, 1, 15,

but it lacks the final sentence-stobha abhrjij jyotir abhrjit.
Oldenberg's examples from the Grmegeyagna.
Oldenberg p. 487, BI. I 420, GG. 193 = :!;tV. 8, 46, 1.
Accentuation
122.2222.
31. 22 . 222 . /
12.22.222.//
Melodization
/1123/auhi
/21.12 3/
/11.223/ -
2.3231/11223/hauhi = 2. 3231 /
/ 1123/
=
I
=
/ 1123/
/hauhi = 2.32312345/. d = 5/
The strophe which should contain 3 x 8 syll. has by contraction got

3 x 7 syll., the same condition is met with in the melodization. The text
is broken up into parvans of 3 or 4 syll. The melodization certainly follows
Oldenberg's rule, if we add to it the secundary rule that vocatives or
parts of vocative expressions which begin a parvan, receive the udtta on
the first two syll.
Oldenberg p. 487, BI. I 443, GG. 209, (not found in the :!;tV.).
Accentuation
12.2.23.122 .
312.23.122/
12.23.1222. / /
Melodization
/55.5.55.555.e = 5/e = 5/
/211. 22 . 1123/hovhi = 1232/
/11.211z3/hovhi = 1232/111232343/
/i = 12345/. d = 5/
80
Examples added to those collected by Oldenberg from the

'Ohyagana (AraJ,lyegeyagana).
Oldenberg has quoted (indirectly) from the AraJ,lyegeyagana 4 examples.
I shall add to them 40thers. The stobhas are only given when they are
of interest.
AG. 1, 2, 25, BI. 11 414. Text UA. 578 = 1;tV. 9, 76, l.
Accentuation
32.31.223.23.23.
12.312.2323.12/
12.232.23.1.123
23.122.222.32.2//
Melodization
/21.21.122.12.12/
/11.211.2212.1'2/
/11.221.12.1.112/
/12.112.222.211. 21 /
Pada d ends in the accented form: nadi{Jv , in the melodized form

nadi{Juv . The strophe theoretically and originally contains 4 x 12 syll.
AG. 2, 1. 29, Jn. 11 430. Text UA. 30 = 1;tV. 7, 32, 22.
Accentuation
31. 2. 22. 223.
122.22.312/
122.31.122.323.
122.22.312//
Melodization
/21.1.22.222 has = 1/
/112.22.211 has = 1/
/112. 2l.112. 211 has = 1/
/112.22.211 has = 1/
AG. 4, 1, 10, BI. 11 468. Text AS. 3, 3 (BI. 11 285) = 1;tV. 1, 91, 2, the
text is a tril;ltubh-strophe. In pada c uru antrik{Jam has been contracted:
u >v, a > and the accent became a kampa, which I have transcribed
by 2"; in two cases tone 3 between two udattas is raised to tone l. The
final sentence-stobha in its accented form runs: jna divam amtrik{Ja1Jt
wtkivt1Jt visvhhojasa1Jt pururp ajijanafl,.
Accentuation
2.3l.122.23.23.
2.31. 2223.2.2. /
l.122. 3.2"123.
l.123.1.12.222//
Melodization
/1.11.112.22.12/
/1.21.1222.1.1/
/1.113.3l.1112/ (sic corr.)
/1.112. 1. 1l.222/
stobha: /11.1l.2112.22l.
21122.2112.2222/
AG. 3, 2, 4, BI. 11 462. Text UA 437 = 1;tV. 1, 4, l. The text is a regular

gayatri without any contraction. However in the melodization the text
is treated very irregularly and is therefore quoted fully.
surpakrtmlm tye
sudUgkm iva goduke/
jukmsi dyvi-dyavi/ /
The fust parvan of the chant-strophe contains the fust p8.da and four
syllables of the second pada with the last syllablei changed into
(i >ai >a 1). The 2d parvan contains the two syllables-va go-; the 3d
81
parvan contans the two text-syllables -dhe plus the sound-stobha uvii;
then there follows the sacral expression ir!Jl and 8var as separate parvans.
The 6th parvan = the first four syllables of pda c; the 7th parvan contains the last tour syllables of this pda folIo wed by the stobha uv; the
last parvan consists of the stobhic expression jyoti~. Melodization:
/22221.212 ;212 .1/222/12.uv = 231113/
/i<J = 3ZM 5/sv1;t = 3l1345/
/211132/1112 uv = 231113/jyotnt = 3l1345/
The influence of the accentuation is clearly traceable, although there
are many irregularities ; the acoustic metalepsis is as a rule ignored, so in
the syllables (u- )tye, (au- )dghm, -dhe; in one case the tone 1 is extended.
over the two following syllables dyvi-dyavi. In the verb juhUmai the
2d syllable receives the udtta. The melody has a regular construction :
the first three parvans move on the higher tones, the 4th and 5th parvans
apply the svrya twice. The 6th and 7th parvans move again on the
higher tones, whilst the last parvan containa again the svrya.
Examples from the Grmegeyagna added to those of Oldenberg.
In the first place I want to consider a sman (GG. 4292) written in the
metre dvipad virj (4 x 5 syllabIes), the first three quintiples of which
are quoted by the commentator N. on PS. 9, 30 as fit for using as stobhas.
The text of this sman (= J;tv. , 9, 109, 4) runs as follows:
pvaava aoma maMnt aamudr~
pita devdnm visvbM dMma/
The melodization runs :

/auhov = 5535/auhov = 2235/auhov auhov = 51i3l134' 55,5/
/112.22/21.121/21 . 2-1212-12/121. ll zM /
/auhov = 55,5/auhov = 2232/auhov auhov = 21 23l134 .55,5/
On the syllable de- of the third verse one would expect tone 2; is the
alteration due to the ornamental character of this parvan 1 Is the s1ight
deviation in the melodization of the introductory and final stobha-lines
caused by a typographical or traditional error?
Tbe two following examples I came across in a cursory perusal of the
Grmegeyagna.
GG. 144' BI. I 339. J;tV. 8, 16. 1.
pr aamrdja1ft car{la!,indm
indra1ft atow nvya1ft girbMM
nra1ft nr{ldha1ft m1fthi{l~ham/ /
a.
b I.
/4.545 . 5554.
45. 55. 4/
82
bIl. /1l.1l.1l.1 23 234 /

c.
/44.545./4 554/ etc.
The chant shows a very ornament al character and yet we can trace in
it the influence of the accentuation. In the tirst place the syllable-division
is very irregular. The first parvan consists of the eight syllables of pda a
plus five syllables of pda b; the next parvan contains the three last
syllable of th is pda; then pda b is repeated in one parvan (b Il); then
there follows pda c divided into 5 + 3 syllabies, whilst the chant-strophe
finishes with parvans containing repeated syllables and stobhas.
The first parvan has a melody-movement in agreement with the accentuation; however there is no acoustic metalepsis, and the udttas are
replaced by tone 4 and the low tones by 5; in the second parvan Mil;, at
the end is lowered from 4 to 5 owing to an ornamental cadence; the 3d
parvan is monotonous on tone 1 with an ornamental descent in the last
two syllabies. In parvan ca the syllables na- and -~- have kept tone 4;
-rmp, shows acoustic metalepsis which is not applied in -harp, since a high
syllable follows; the last syllable of parvan cf3 is raised, cf. the irregularity
at the end of the pda bI.
We see that influence of accentuation and sacral monotony are melodie
principles which guide but do not bind.
GG. 2513, BI. I 519, l;{,V. 8, 3, 15.
ud u ty mdhumattam
gira(lJ) stmsa irate
satrjito dhanasd k~itotayo
vjaynto rth iva.
/ha>p. = 32345 /3.4.3. 345/33 234hi = 5/
/112.112/3.2345/32345/
c.
/2212-12' 221. 112232345 /hazp = 32345 /
d I /4443.45./33 234 Mi = 5/
dIl /2212. 12.2/l./auhov = 2345/hi = 4sN = 5/
a.
b.
Here again the influence of the accentuation, although hidden, is

traceable. In the parvan consisting of the syllables ud u tye madhuma- the
udttas are represented by tone 3; the last parvan of pada a is ornamental
and contains a des cent from 3 to 5, which is a typical motive for the
chant. Parvan ba has a partial acoustic metalepsis by which the syllables
-ra(lJ) and -m- become 12 and 12, In pda c -ji-, -s and a- have kept
tone 1, the syllables -to and -k~i- have received acoustic metalepsis. The
first parvan of d I goes against the spoken accentuation - read: 4433. 45 1
- and the second parvan of d I has the typical descent ; on the other
hand parvan a of pda d Il with seven syllables follows the spoken
accentuation without acoustic metalepsi~; -va lengthened into -v has
received tone 1 as is often the case in such a parvan formed by one syll.
In my opinion the study of the Smaveda has to be accompanied by a
83
detailed interpretation of the Pu~pastra . Now, at the close of this stage

of my investigation I want to express my admiration for Hermann
Oldenberg, whose vivid intuition has laid the basis for the musicological
study of the Smaveda.
POSTSCRIPT
The manuscript in handwriting being ready in October 1948, the copying

in type-writing took me owing to a serious illness nearly two years.

Studies On The Samaveda - Faddegon, B PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Studies On The Samaveda - Faddegon, B PDF

Hochgeladen von

Copyright:

Verfügbare Formate

STUDIES ON THE SMAVEDA

VERHANDELINGEN DER KONINKLIJKE NEDERLANDSE

NIEUWE REEKS DEEL LVII, N. 1

SECOND PRINTING: 1963

at a meeting of tke Koninklijke Nederlandse Akademie van

The edition of the Sm.veda-Sarphit, 9; publications by R. Simon and

J. M. van der Hoogt, 9.

The Smavedic gamut according to Burnell and Se!i!&giri Sstri.

STUDIES ON THE SMAVEDA

The general line in the development of music

THREE PRELIMINARY NOTES

1. The indication of the changes in the text when melodized, 35;

Fox Strangways' enthusiasm and ina.ccura.cy, 49; information given by

Olassijication of the chants of the Griimegeyagiina .

J. General remarks; 11. Scheme of subdivision in the following

A. The final parvan consists of text..:syllables or a sentence-stobha; a. the

GENERAL CONCLUSION TO THE SIX SUBSECTIONS

Function of the different tones of the gamut as final tone

Analysis of the smans ending in 5656

Cla.ssification in tabellic form, p. 70.

Subgroup 111 /323455856/' . . .

Cla.ssification in tabellic form, p. 84.

Oldenberg's rule of acoustic metalepsis. . .

Oldenberg's examples from the AraQyegeyagna (Ohyagna), 78; from the

STUDIES ON THE SMAVEDA

in fixed proportion of time; as the use of this gracenote is bound to the